The license-api module was added in 858975f0 but the CI deploy step's
`-pl` list still only built parent + server-core + minter. server-core
now depends on cameleer-license-api, which wasn't in the registry yet,
so the deploy job failed with:
Could not find artifact com.cameleer:cameleer-license-api:jar:1.0-SNAPSHOT
in gitea (https://gitea.siegeln.net/api/packages/cameleer/maven)
Add cameleer-license-api to the project list so it builds and publishes
before its consumers in the same reactor invocation.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Splits the pure license contract types (LicenseInfo, LicenseValidator,
LicenseState, LicenseStateMachine, LicenseLimits, DefaultTierLimits) into a
new cameleer-license-api module under package com.cameleer.license.
Why: cameleer-license-minter previously depended on cameleer-server-core for
these types, dragging cameleer-server-core + cameleer-common onto the
classpath of every minter consumer (notably cameleer-saas). The SaaS
management plane has no business carrying server-runtime types — it only
needs the license contract to mint and verify tokens.
After:
cameleer-license-minter -> cameleer-license-api (no server internals)
cameleer-server-core -> cameleer-license-api
cameleer-saas -> cameleer-license-minter -> cameleer-license-api
Verified: mvn -pl cameleer-license-minter dependency:tree shows the minter
no longer pulls cameleer-server-core or cameleer-common. Full reactor
verify (-DskipITs) green: 371 tests pass.
LicenseGate stays in server-core (server-runtime state holder, not contract).
Closescameleer/cameleer-server#156
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the prompt=none → /login?local trap with a deterministic
capability endpoint (GET /api/v1/auth/capabilities). LoginPage renders
SSO-primary or local form based on caps; ?local is the explicit
admin-recovery escape hatch. Drops prompt=none from the SSO authorize
URL per RFC 9700 §4.4. Adds Vitest + IT coverage and docs.
MFA enrollment / enforcement deferred to issue #154.
Adds <distributionManagement> at the parent POM and a push-only deploy
step in the build job. Selects the parent + core + minter via -pl so
both the plain library JAR and the Spring Boot fat CLI JAR are pushed
with their full dep tree resolvable; server-app is excluded as a
fat-jar runtime, not a library.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Captures the decision to gate login UX on capabilities (no SaaS-mode
flag), drop prompt=none from the primary OIDC flow per RFC 9700 §4.4,
and keep ?local as the explicit admin-recovery escape hatch.
MFA enrollment / enforcement and password reset for local accounts are
explicitly deferred and tracked in issue #154.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Final consolidation pass after the 36-task license-enforcement work.
core-classes.md:
- New license/ section: LicenseInfo, LicenseLimits, DefaultTierLimits,
LicenseValidator, LicenseGate, LicenseStateMachine, LicenseState.
- runtime/: added CreateGuard (functional interface for license-cap
hooks consulted by EnvironmentService/AppService/AgentRegistryService).
- admin/: AuditCategory.LICENSE added to the documented enum value list.
app-classes.md:
- New license/ section: LicenseService, LicenseRepository, LicenseRecord,
PostgresLicenseRepository, LicenseChangedEvent, LicenseEnforcer,
LicenseUsageReader, LicenseCapExceededException, LicenseExceptionAdvice,
LicenseMessageRenderer, RetentionPolicyApplier, LicenseRevalidationJob,
LicenseMetrics.
- LicenseAdminController entry expanded to document the GET response
shape and the LicenseService.install delegation pattern.
- config/: RuntimeBeanConfig note about CreateGuard wiring; new
LicenseBeanConfig entry covering the four-bean topology and the
always-failing-validator fallback.
Note: LicenseChangedEvent, LicenseRepository, LicenseRecord, and
PostgresLicenseRepository live in cameleer-server-app, not -core; the
plan's section assignments were corrected against the actual code.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two new assertions: license table has tenant_id/license_id/token/
installed_at/installed_by/expires_at/last_validated_at columns with
expected types + NOT NULL constraints, PK on tenant_id; environments
has execution_retention_days/log_retention_days/metric_retention_days
all integer NOT NULL DEFAULT 1.
Note: V5 migration does not include an installed_via column; the
plan's spec was aspirational. Test asserts what the migration
actually creates (and what PostgresLicenseRepository reads/writes).
OpenAPI regen (Step 35.2) deferred to session end — requires running
backend + UI dev server.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Install license with max_log_retention_days=30, env.configured=60 →
effective=30; verify ClickHouse logs table reflects toIntervalDay(30).
Replace with max=7 → effective=7; verify TTL recomputed. Polls
system.tables.create_table_query up to 5s for the @Async listener
to apply.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Five @Nested cap surfaces (envs, apps, outbound, alert rules, users)
share a single synthetic license with cap=1 each. Each test pushes
just past the cap and verifies the standard 403 envelope plus a
cap_exceeded audit row. Per-limit ITs cover full per-cap behavior;
this IT catches accidental wire-rip regressions across all caps.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
End-to-end IT covering the full lifecycle: mint a token via
cameleer-license-minter (test-scope), POST it via /api/v1/admin/license,
verify state=ACTIVE, clear gate, revalidate from PG, verify state restored.
Plus: tampered signature -> 400 + LICENSE/FAILURE audit row, gate not
mutated to ACTIVE.
Adds cameleer-license-minter as a test-scope dep on cameleer-server-app
(verified absent from runtime/compile classpaths). Also disables the
default spring-boot:repackage execution on the minter pom so the main
artifact stays as a plain library JAR consumable as a Maven dependency
(the cli classifier still produces the executable jar).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
cameleer_license_state{state=...} (one-hot per LicenseState),
cameleer_license_days_remaining (negative when ABSENT/INVALID),
cameleer_license_last_validated_age_seconds. Refreshed on
LicenseChangedEvent and every 60s.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Returns state, expiresAt/daysRemaining, lastValidatedAt, message
(LicenseMessageRenderer.forState), and a limits[] array where each
entry carries key/current/cap/source ("license" vs "default"). Adds
public AgentRegistryService.liveCount() so max_agents can be reported
from the in-memory registry.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
GET returns {state, invalidReason, envelope, lastValidatedAt}. POST
delegates to licenseService.install(token, userId, "api") so install
goes through audit + persistence + event publish. Removes the inline
LicenseValidator construction from the controller.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@Scheduled(cron = "0 0 3 * * *") triggers svc.revalidate() daily.
@EventListener(ApplicationReadyEvent.class) @Async fires once 60s
after boot to catch ABSENT->ACTIVE transitions if the license was
written to PG between server starts. Exceptions are logged but never
propagate.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@EventListener fires on every license install/replace/expire. For each
environment, computes effective TTL = min(licenseCap, env.configured)
and emits one ALTER TABLE ... MODIFY TTL ... per (table, env). Tables
covered: executions, processor_executions, logs, agent_metrics,
agent_events. ClickHouse failures are logged but do not propagate
(listener is async-tolerant).
route_diagrams is intentionally excluded -- it has no TTL clause in
init.sql (ReplacingMergeTree keyed on content_hash, not time-series).
server_metrics is also excluded -- it has no environment column
(server straddles environments).
Per-environment TTL via WHERE requires ClickHouse 22.3+; the project's
current image (clickhouse/clickhouse-server:24.12) is well above that
floor.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds three int fields to the Environment record + repository row mapper,
matching the columns added in V5. Default value is 1 per the V5 NOT NULL
DEFAULT 1. Read DTO surfaces the fields via Jackson record serialization;
setter endpoint deferred to a follow-up that wires the corresponding
license cap checks.
The canonical constructor enforces >= 1 for each retention field — V5
guarantees this at the DB level, but the runtime guard catches in-memory
construction errors (e.g., test sites that pass 0).
Test sites updated to the 12-arg signature with retention defaults of 1.
EnvironmentAdminControllerIT gains a regression test asserting the wire
shape exposes all three fields.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Returns 422 UNPROCESSABLE_ENTITY when jarRetentionCount exceeds
license cap. Default tier cap = 3. The other three retention caps
(execution/log/metric retention days) are deferred to T26+ where
the corresponding fields are added to Environment.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds ComputeUsage record + computeUsage() helper to LicenseUsageReader
that aggregates from PG. DeploymentExecutor.executeAsync runs three
assertWithinCap checks (max_total_cpu_millis, max_total_memory_mb,
max_total_replicas) right after config resolution. The existing
executor try/catch turns a LicenseCapExceededException into a FAILED
deployment with the cap message in the failure reason.
Adds ComputeCapEnforcementIT (HTTP-driven; @MockBean RuntimeOrchestrator,
since cap rejection short-circuits before any orchestrator call) plus
defensive license lifts in BlueGreenStrategyIT, RollingStrategyIT,
DeploymentSnapshotIT, and DeploymentControllerAuditIT so sequential
deploys under testcontainer reuse don't trip the new caps.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds AlertRuleRepository.count() and a LicenseEnforcer.assertWithinCap
call at the top of the POST handler. Default cap = 2; the 3rd rule
gets the standard 403 envelope. Sibling alert ITs that legitimately
need more than 2 rules get the cap lifted via the test-license helper.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds LicenseEnforcer.assertWithinCap call at the top of create() using
repo.listByTenant(tenantId).size() as the current count. Lifts the cap
in OutboundConnectionAdminControllerIT (duplicateNameReturns409 needs
2 creates in one test). LicenseExceptionAdvice maps the rejection to
the standard 403 envelope; cap_exceeded audit row emitted via the
LicenseEnforcer 3-arg ctor.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wires LicenseEnforcer into UserAdminController.createUser and
OidcAuthController auto-signup. Cap fires before any validation so
over-cap creates short-circuit cheaply. Audit emission already
present (LicenseEnforcer 3-arg ctor from T16 emits cap_exceeded
under AuditCategory.LICENSE).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a CreateGuard to AgentRegistryService that fires only on NEW
registrations: re-registers of an existing agent bypass the cap (they
don't grow the registry, and rejecting them would orphan an agent that
already counts against the cap). Live-only count for cap enforcement —
STALE/DEAD/SHUTDOWN agents are excluded so the cap reflects the working
fleet, not historical residue.
Reuses the CreateGuard pattern from T18-T19. The global
LicenseExceptionAdvice maps the resulting LicenseCapExceededException to
403 with the structured envelope — no AgentRegistrationController
changes needed.
AgentCapEnforcementIT exercises the HTTP path end-to-end: two registers
succeed at cap=2, a third returns 403 with the expected envelope, and a
re-register of an already-registered agent succeeds at-cap.
Sibling agent-registering ITs (Agent*ControllerIT, Diagram*IT,
Execution*IT, Search*IT, Protocol*IT, Backpressure*IT, JwtRefresh*IT,
Registration*IT, Security*IT, SseSigning*IT, IngestionSchemaIT) lift
max_agents in @BeforeEach and clear the synthetic license in @AfterEach
— the in-memory registry is shared across @SpringBootTest reuse
boundaries, so without the lift the default-tier max_agents=5 would be
exhausted by accumulated test residue.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds CreateGuard hook to AppService.createApp using the same pattern
as T18 (EnvironmentService). AppRepository.count() added; the bean
wires LicenseEnforcer.assertWithinCap("max_apps", current, 1).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Makes the signature-bypass loud at every call site since T19-T25 will
copy this pattern 5+ more times. The helper still loads via
LicenseGate.load() directly (no signature check) — the new name
ensures any future caller has to acknowledge that.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds CreateGuard functional interface to core (preserves the no-Spring
boundary between core and app) and wires LicenseEnforcer into the
EnvironmentService bean in RuntimeBeanConfig so POST
/api/v1/admin/environments rejects with the structured 403 envelope
(error/limit/cap/state/message) once the cap is reached. Default tier
max_environments=1; the V1 baseline seeds the default env, so the very
next create through the API is rejected unless a license lifts the cap.
Also adds EnvironmentRepository.count() (with PostgresEnvironmentRepository
impl), TestSecurityHelper.installTestLicenseWithCaps(...) so existing ITs
that POST envs keep working, and a defensive cleanup in
LicenseUsageReaderIT/EnvironmentAdminControllerIT to stay
order-independent under Testcontainer reuse (deletes deployments+apps
before envs to avoid FK violations).
Test: EnvironmentCapEnforcementIT (new) drives the rejection path
end-to-end and asserts the 403 body shape produced by
LicenseExceptionAdvice.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
One COUNT per entity table; one SUM-grouped query over non-stopped
deployments for compute caps. SQL traverses
deployed_config_snapshot->'containerConfig' (corrected from the
plan's top-level path; the snapshot record nests containerConfig
under that key). agentCount is fed in by the controller since it's
an in-memory registry value, not a DB row.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two follow-ups to LicenseEnforcer review:
- Add @Autowired to the 3-arg ctor so Spring picks it unambiguously
(the 2-arg test ctor is otherwise an equally-greedy candidate).
- Wrap audit.log() in try/catch + log.warn so a degraded audit DB
cannot mask a cap rejection: callers still see HTTP 403 even when
audit storage is unhealthy.
- Extract counter name to private static final.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
assertWithinCap consults LicenseGate.getEffectiveLimits, throws
LicenseCapExceededException on overflow, increments
cameleer_license_cap_rejections_total{limit=...} for telemetry, and
emits an AuditCategory.LICENSE cap_exceeded audit row when an
AuditService is wired (3-arg ctor; the test-only 2-arg ctor passes
null and the audit call short-circuits). Unknown limit keys are
programmer errors (IllegalArgumentException), not 403s.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
When username is empty (e.g. email-registered OIDC users with no display
name), the badge was hidden entirely, making logout inaccessible. Always
render the badge with fallback text "Account".
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
LicenseCapExceededException + @ControllerAdvice mapping to 403 with a
body that includes state, limit, current, cap, and a per-state human
message templated by LicenseMessageRenderer (covers ABSENT/ACTIVE/
GRACE/EXPIRED/INVALID with day counts and reason). Adds the forState()
overload now (used by the /usage endpoint in Task 30) so both surfaces
share identical phrasing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Resume point for the next session executing the License Enforcement
plan. Captures: 14 done commit SHAs, what works/doesn't end-to-end,
critical plan deviations (AuditService.log API; LicenseInfo.label
not tier; throwaway-keypair fallback validator; ClickHouse TTL WHERE
caveat for T27), batching strategy, and suggested next-task order.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LicenseBootLoader @PostConstruct calls LicenseService.loadInitial,
which delegates to install() so env-var/file/DB paths share a single
audit + event-publish code path. A missing public key now produces
an always-failing validator (constructed with a throwaway keypair so
the parent ctor accepts it) so loaded tokens route to INVALID
instead of being silently ignored.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Single mediation point for token install/replace/revalidate. Audits
under AuditCategory.LICENSE, persists to PG, mutates the LicenseGate,
and publishes LicenseChangedEvent so downstream listeners
(RetentionPolicyApplier, LicenseMetrics) react uniformly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
JdbcTemplate-backed repo; upsert is ON CONFLICT (tenant_id), touch
updates only last_validated_at, delete is provided for future
operator-clear flow (not exposed as REST in v1).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Per-tenant license row stores the signed token, licenseId for audit,
installed/expires/last_validated timestamps. environments gains three
INTEGER NOT NULL DEFAULT 1 retention columns (execution, log, metric)
so existing rows land inside the default-tier cap.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds --verify (requires --public-key) to LicenseMinterCli. After
writing the output file the CLI parses the freshly-minted token
through LicenseValidator against the supplied public key. On
verify failure the output file is deleted (so the bad token is
not accidentally shipped) and the CLI exits 3.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reads PEM or base64 PKCS#8 Ed25519 private key, maps --max-foo-bar
flags to max_foo_bar limit keys, parses --expires as a UTC date,
defaults --grace-days to 0. Unknown flags fail fast with exit 2.
--verify path is added in the next task.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pure signing primitive: serialises LicenseInfo to canonical JSON
(sorted top-level keys via ORDER_MAP_ENTRIES_BY_KEYS plus a TreeMap
for the limits sub-object) then signs with Ed25519. Round-trips
through LicenseValidator and is byte-stable across runs for
identical inputs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Top-level module sibling to cameleer-server-core/-app. Depends on
cameleer-server-core for the LicenseInfo schema. Spring Boot
repackage produces a runnable -cli classifier for the vendor.
Not added as a dependency from cameleer-server-app — runtime tree
must not carry signing primitives.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LicenseGate now exposes getState() (delegates to LicenseStateMachine),
getEffectiveLimits() (merged over DefaultTierLimits in ACTIVE/GRACE,
defaults-only in ABSENT/EXPIRED/INVALID), markInvalid(reason), and
clear(). Internal snapshot is an immutable record-like class swapped
atomically so concurrent reads see a consistent license+reason pair.
Removes the transient openSentinel() and getTier() introduced by
earlier tasks (no production consumers).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pure-domain FSM (ABSENT/ACTIVE/GRACE/EXPIRED/INVALID) and the
default-tier constants per spec §3. invalidReason wins over any
loaded license so signature failures surface as INVALID rather
than masking as ABSENT.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec §2.1 — both fields are required and the validator rejects a
token whose tenantId does not match the server's configured tenant
(CAMELEER_SERVER_TENANT_ID). Self-hosted customers cannot strip
tenantId because the field is in the signed payload.
LicenseBeanConfig and LicenseAdminController updated to pass the
expected tenant to the validator constructor. The transient
placeholder/TODO from Task 2 is removed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Required fields per spec §2.1. tenantId is non-blank; gracePeriodDays
defines the post-exp window during which limits keep applying.
isExpired() now honours the grace; isAfterRawExpiry() distinguishes
ACTIVE from GRACE for the state machine in Task 4.
Validator and gate use placeholder values temporarily; Task 3 wires
the validator to read the new fields, Task 5 rewrites the gate.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec §9 — feature flags are out of scope for license enforcement.
Drops Feature.java, LicenseGate.isEnabled, LicenseInfo.hasFeature,
and the corresponding test cases. LicenseValidator now silently
ignores any features array on the wire (no error).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add INVALID state to FSM (signature/tenant/parse failure ≠ ABSENT)
with loud UI/audit/metric severity; ABSENT stays a calm state.
- Make tenantId required in the license envelope (it's already inside
the signed payload, so a self-hosted customer cannot strip it).
- Move ClickHouse TTL recompute from boot-only to a
RetentionPolicyApplier @EventListener(LicenseChangedEvent), so a
long-running server that lands in EXPIRED tightens TTL automatically.
- Add LicenseRevalidationJob (daily) that re-runs signature check
against the DB row and updates last_validated_at; transitions to
INVALID on failure (catches public-key rotation drift).
- Add last_validated_at column to the license table, surfaced on the
/usage endpoint and as cameleer_license_last_validated_age_seconds.
- Enrich enforcement-failure responses and the /usage endpoint with a
per-state human-readable message so 403s and the UI both explain
WHY caps changed.
- Add --verify (with --public-key) to the minter CLI to round-trip a
freshly-minted token through LicenseValidator before shipping it,
deleting the output file on verify failure.
- Add corresponding tests, telemetry gauge, and a runtime-recompute IT.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Captures the agreed design for enforcing licensing on cameleer-server:
- Default tier with hard caps when no license is configured
- Arbitrary per-customer limits in signed Ed25519 license tokens
- Standalone cameleer-license-minter module (vendor-only)
- DB-persisted license with env/file override paths
- ABSENT/ACTIVE/GRACE/EXPIRED state machine; offline expiry only
- Removes the dead Feature enum scaffolding
Pending writing-plans.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Surfaces the multi-tenant container hardening contract introduced in the
prior commit so operators and integrators know what is enforced and why.
- application.yml: declare `cameleer.server.runtime.dockerruntime`
alongside the other runtime properties (empty = auto-detect runsc).
- HOWTO.md: add the override row to the Runtime config table.
- SERVER-CAPABILITIES.md: new "Multi-Tenant Runtime Sandboxing" section
describing the cap_drop, no-new-privileges, AppArmor, read-only rootfs,
pids_limit, /tmp tmpfs, and runsc auto-detect contract — plus the
on-disk state caveat that motivates issue #153.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Tenant JARs are arbitrary user code: Camel ships components (camel-exec,
camel-bean, MVEL/Groovy templating) that turn a header into shell, and
Java 17 has no SecurityManager — the JVM is not a security boundary.
This applies an unconditional hardening contract to every tenant
container so a single runc CVE no longer equals host takeover.
DockerRuntimeOrchestrator.startContainer now sets:
- cap_drop ALL (Capability.values() — docker-java has no ALL constant)
- security_opt: no-new-privileges, apparmor=docker-default
(default seccomp profile applies implicitly)
- read_only rootfs, pids_limit=512
- /tmp tmpfs rw,nosuid,size=256m — no noexec, since Netty/Snappy/LZ4/Zstd
dlopen native libs from /tmp via mmap(PROT_EXEC) which noexec blocks
The orchestrator also probes `docker info` at construction and uses runsc
(gVisor) automatically when the daemon has it registered. Override via
cameleer.server.runtime.dockerruntime (e.g. "kata"); empty = auto.
Outbound TCP, DNS, and TLS are unaffected — caps/seccomp don't gate
those — so vanilla Camel-Kafka producers/consumers and REST integrations
keep working unchanged. Stateful tenants (Kafka Streams with on-disk
state stores, apps writing to /var/log/...) need explicit writeable
volumes; that's tracked in #153 as the natural follow-up.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Live-pushed config fields (taps, tapVersion, tracedProcessors,
routeRecording) apply via SSE CONFIG_UPDATE — they take effect on
running agents without a redeploy and are fetched on agent restart
from application_config. They must not contribute to the
"pending deploy" diff against the last-successful-deployment snapshot.
Before this fix, applying a tap from the process diagram correctly
rolled out in real time but then marked the app "Pending Deploy (1)"
because DirtyStateCalculator compared every agentConfig field. This
also contradicted the UI rule (ui.md) that the live tabs "never mark
dirty".
Adds taps, tapVersion, tracedProcessors, routeRecording to
AGENT_CONFIG_IGNORED_KEYS. Updates the nested-path test to use a
staged field (sensitiveKeys) and adds a new test asserting that
divergent live-push fields keep dirty=false.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add a StatusDot + colored Badge next to the app name in the deployment
page header, showing the latest deployment's status (RUNNING / STARTING
/ FAILED / STOPPED / DEGRADED / STOPPING). The existing "Pending
deploy" badge now carries a tooltip explaining *why*: either a list of
local unsaved edits, or a per-field diff against the last successful
deploy's snapshot (field, staged vs deployed values). When server-side
differences exist, the badge shows the count.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add a repeatable attr query parameter to the GET /executions endpoint that
parses key-only (exists check) and key:value (exact or wildcard-via-*)
filters. Invalid keys are mapped to HTTP 400 via ResponseStatusException.
The POST /executions/search path already honoured attributeFilters from
the request body via the Jackson canonical ctor; an IT now proves it.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Drop the page-local DS Select window picker. Drive from() / to() off
useGlobalFilters().timeRange so the dashboard tracks the same TopBar range
as Exchanges / Dashboard / Runtime. Bucket size auto-scales via
stepSecondsFor(windowSeconds) (10 s for ≤30 min → 1 h for >48 h). Query
hooks now take ServerMetricsRange = { from: Date; to: Date } instead of a
windowSeconds number, so they support arbitrary absolute or rolling ranges
the TopBar may supply (not just "now − N"). Toolbar collapses to just the
server-instance badges.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
SERVER-CAPABILITIES.md now lists the two consumption paths (UI + REST API)
side-by-side with visibility rules; the dashboard-builder doc leads with a
"Built-in admin dashboard" section and a 2026-04-24 changelog entry so
first-time readers know they don't have to build anything before seeing
server health.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds /admin/server-metrics page mirroring the Database/ClickHouse visibility
rules: sidebar entry gated on capabilities.infrastructureEndpoints, backend
controller now has @ConditionalOnProperty(infrastructureendpoints) and
class-level @PreAuthorize('hasRole(ADMIN)'). Dashboard panels are driven
from docs/server-self-metrics.md via the generic
/api/v1/admin/server-metrics/{catalog,instances,query} API — Server Health,
JVM, HTTP & DB pools, and conditionally Alerting + Deployments when their
metrics appear in the catalog. ThemedChart / Line / Area from the design
system; hooks in ui/src/api/queries/admin/serverMetrics.ts. Not yet
browser-verified against a running dev server — backend IT covers the API
end-to-end (8 tests), UI typecheck + production bundle both clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds /api/v1/admin/server-metrics/{catalog,instances,query} so SaaS control
planes can build the server-health dashboard without direct ClickHouse
access. One generic /query endpoint covers every panel in the
server-self-metrics doc: aggregation (avg/sum/max/min/latest), group-by-tag,
filter-by-tag, counter-delta mode with per-server_instance_id rotation
handling, and a derived 'mean' statistic for timers. Regex-validated
identifiers, parameterised literals, 31-day range cap, 500-series response
cap. ADMIN-only via the existing /api/v1/admin/** RBAC gate. Docs updated:
all 17 suggested panels now expressed as single-endpoint queries.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Snapshot the full Micrometer registry (cameleer business metrics, alerting
metrics, and Spring Boot Actuator defaults) every 60s into a new
server_metrics table so server health survives restarts without an external
Prometheus. Includes a dashboard-builder reference for the SaaS team.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Second analyze pass after pushing showed a slightly different symbol
count. Counts-only bump.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Auto-bumped by `npx gitnexus analyze --embeddings` after the diagram
refactor landed. No content changes — counts only.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
useCatalog now accepts optional from/to query params and LayoutShell
threads the TopBar time range through, so the per-app exchange counts
shown in the sidebar align with the Exchanges tab window. Previously
the sidebar relied on the backend's 24h default — 73.5k in the sidebar
coexisted with 0 hits in a 1h Exchanges search, confusing users.
Other useCatalog callers stay on the default (no time range), matching
their existing behavior.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
handleLogout explicitly cleared the env from localStorage, forcing the
env switcher modal to re-open on every login. Drop that clear so the
last selected env is restored from localStorage on the next session —
the expected behavior for a personal-preference store.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- app-classes: DiagramRenderController by-route endpoint no longer
depends on the agent registry; points at findLatestContentHashForAppRoute
and cross-refs the exchange viewer's content-hash path.
- core-classes: document the new DiagramStore method and note why the
agent-scoped findContentHashForRoute stays for the ingest path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Store-level: assert findLatestContentHashForAppRoute picks the newest
hash across publishing instances (proves the lookup survives agent
removal), isolates by (app, env), and returns empty for blank inputs.
Controller-level: assert the env-scoped /routes/{routeId}/diagram
endpoint resolves without a registry prerequisite, 404s for unknown
routes, and that an execution's stored diagramContentHash stays pinned
to the point-in-time version after a newer diagram is stored — the
"latest" endpoint flips to v2, the by-hash render remains byte-stable.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
All production callers migrated to findLatestContentHashForAppRoute in
the preceding commits. The agent-scoped lookup adds no coverage beyond
the latest-per-(app,env,route) resolver, so the dead API is removed
along with its test coverage and unused imports.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Both catalog controllers resolved the from-endpoint URI via
findContentHashForRouteByAgents, which filtered by the currently-live
agent instance_ids. Routes removed between app versions therefore lost
their fromUri even though the diagram row still exists.
Route through findLatestContentHashForAppRoute so resolution depends
only on (app, env, route) — stays populated for historical routes.
CatalogController now resolves the per-row env slug up-front so the
fromUri lookup works even for cross-env queries against managed apps.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The env-scoped /routes/{routeId}/diagram endpoint filtered diagrams by
the currently-live agent instance_ids. Routes removed between app
versions have no live publisher, so the lookup returned 404 even though
the historical diagram row still exists in route_diagrams. Sidebar
entries for removed routes showed "no diagram" as a result.
Switch to findLatestContentHashForAppRoute which resolves directly off
(applicationId, environment, routeId) + created_at DESC, independent of
the agent registry. The controller no longer depends on
AgentRegistryService.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Agent-scoped lookups miss diagrams from routes whose publishing agents
have been redeployed or removed. The new method resolves by
(applicationId, environment, routeId) + created_at DESC, independent of
the agent registry. An in-memory cache mirrors the existing hashCache
pattern, warm-loaded at startup via argMax.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Added the new Traefik TLS cert resolver setting to the runtime env var
table. Blank default matches how ACME-less dev/local installs want the
`tls.certresolver` label omitted entirely.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previously `TraefikLabelBuilder` hardcoded `tls.certresolver=default` on
every router. That assumes a resolver literally named `default` exists
in the Traefik static config — true for ACME-backed installs, false for
dev/local installs that use a file-based TLS store. Traefik logs
"Router uses a nonexistent certificate resolver" for the bogus resolver
on every managed app, and any future attempt to define a differently-
named real resolver would silently skip these routers.
Server-wide setting via `CAMELEER_SERVER_RUNTIME_CERTRESOLVER` (empty by
default) flows through `ConfigMerger.GlobalRuntimeDefaults.certResolver`
into `ResolvedContainerConfig.certResolver`. When blank the
`tls.certresolver` label is omitted entirely; `tls=true` is still
emitted so Traefik serves the default TLS-store cert. When set, the
label is emitted with the configured resolver name.
Not per-app/per-env configurable: there is one Traefik per server
instance and one resolver config; app-level override would only let
users break their own routers.
TDD: TraefikLabelBuilderTest gains 3 cases (resolver set, null, blank).
Full unit suite 211/0/0.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds a boolean `externalRouting` flag (default `true`) on
ResolvedContainerConfig. When `false`, TraefikLabelBuilder emits only
the identity labels (`managed-by`, `cameleer.*`) and skips every
`traefik.*` label, so the container is not published by Traefik.
Sibling containers on `cameleer-traefik` / `cameleer-env-{tenant}-{env}`
can still reach it via Docker DNS on whatever port the app listens on.
TDD: new TraefikLabelBuilderTest covers enabled (default labels present),
disabled (zero traefik.* labels), and disabled (identity labels retained)
cases. Full module unit suite: 208/0/0.
Plumbed through ConfigMerger read, DeploymentExecutor snapshot, UI form
state, Resources tab toggle, POST payload, and snapshot-to-form mapping.
Rule files updated.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The field was cosmetic — `containerConfig.exposedPorts` only fed Docker's
`Config.ExposedPorts` metadata via `withExposedPorts(...)`. It never
published a host port and Traefik routing uses `appPort` from the label
builder, not this list. Users reading the label "Exposed Ports" reasonably
expected it to expose their port externally; removing it until real
multi-port Traefik routing lands (tracked in #149).
Backend DTOs (`ContainerRequest.exposedPorts`, `ConfigMerger.intList
("exposedPorts")`) are left in place so existing containerConfig JSONB
rows continue to deserialize. New writes from the UI will no longer
include the field.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The exclusion list still named the legacy flat `/api/v1/search/executions`
URL, which no longer exists — the endpoint moved to env-scoped
`/api/v1/environments/{envSlug}/executions/search`. Exact-match Set
lookup never matched, so every UI search POST produced an audit row.
Switch to AntPathMatcher over a pattern list so the dynamic envSlug is
handled correctly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Surface from the Task 0 testcontainers.reuse enable: when the same Postgres
container is reused across `mvn verify` runs, Flyway migrates both `public`
and `tenant_default` schemas (the app.yml default URL uses
?currentSchema=tenant_default; AbstractPostgresIT overrides to public).
Schema-introspection assertions saw duplicate rows/indexes/enums.
Plus: OutboundConnectionAdminControllerIT's AfterEach couldn't delete its
test users because sibling deployment ITs (Task 4) left deployments.created_by
references — FK blocks the DELETE. Clear referencing deployments first.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Extract inline handleRestore mapping into snapshotToForm(snapshot, defaults) helper
- Export defaultForm from useDeploymentPageState for use in ConfigPanel
- Replace ConfigPanel stub with real read-only snapshot renderer + Snapshot/Diff toggle
- Add fieldDiff deep-equal field-walk helper with nested object + array support
- Forward optional currentForm prop through CheckpointDetailDrawer to ConfigPanel
- 13 new tests across diff.test.ts, snapshotToForm.test.ts, ConfigPanel.test.tsx (all pass)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extract 14 inline style blocks from CheckpointDetailDrawer index.tsx and
LogsPanel.tsx into a shared CSS module using DS CSS variables throughout.
Narrow the LogsPanel useMemo dep array from the full deployment object to
deployment.id + deployment.replicaStates to prevent spurious query
invalidation on every TanStack Query poll.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds the CheckpointDetailDrawer with Logs/Config tabs. LogsPanel scopes
logs to a deployment's replicas via instanceIds derived from replicaStates
+ generation suffix. Stub ConfigPanel placeholder for Task 11.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Fetched fresh openapi.json from local backend (Tasks 3-5 changes)
- Regenerated schema.d.ts via openapi-typescript
- Added createdBy: string | null to Deployment interface in apps.ts
- Added instanceIds?: string[] to UseInfiniteApplicationLogsArgs with sort/serialize/queryKey/URLSearchParams wiring
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds List<String> instanceIds to LogSearchRequest (null-normalized to
List.of() in compact ctor) and generates an IN clause in both
ClickHouseLogStore.search() and countLogs(), mirroring the existing
sources pattern. LogQueryController parses ?instanceIds= as a
comma-split list. All existing LogSearchRequest call sites updated.
New ClickHouseLogStoreInstanceIdsIT covers: multi-value filter, empty
filter (all rows), null filter (all rows), single-value filter, and
coexistence with the singular instanceId field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Wires AuditService and AppVersionRepository into DeploymentController.
Replaces null createdBy placeholder with currentUserId() on createDeployment/promote.
Adds audit log entries (SUCCESS + FAILURE) for deploy_app, stop_deployment,
and promote_deployment actions. Fixes FK violations in affected ITs by
seeding the test-operator and alice users into the users table before deploy calls.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Fix Issue 1: Add @AfterEach cleanup for alice/bob users in PostgresDeploymentRepositoryCreatedByIT to prevent test leakage (FK order: deployments -> app_versions -> apps, then users).
Fix Issue 2: Add comment at first create(..., null) call site in PostgresDeploymentRepositoryIT documenting the null placeholder for pre-V4 rows where createdBy is nullable.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Appends String createdBy to the Deployment record (after createdAt), updates
both with-er methods to pass it through, threads the parameter through
DeploymentRepository.create, DeploymentService.createDeployment/promote, and
PostgresDeploymentRepository (INSERT + SELECT_COLS + mapRow). DeploymentController
passes null as placeholder (Task 4 will resolve from SecurityContextHolder).
Covers with PostgresDeploymentRepositoryCreatedByIT verifying round-trip via
both createDeployment and promote.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
15 tasks across 5 phases (backend foundation → SideDrawer →
ConfigTabs readOnly → CheckpointsTable + DetailDrawer → polish).
TDD throughout with per-task commits. Backend phase ships
independently to close the audit gap as quickly as possible.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the cramped Checkpoints disclosure with a real DataTable + a
side drawer (Logs / Config with snapshot/diff modes) and closes the
audit-log gap discovered in DeploymentController (deploy/stop/promote
currently make zero auditService.log calls).
Cap visible checkpoints at Environment.jarRetentionCount — beyond that,
JARs are pruned and rows aren't restorable. Logs scoped per-deployment
via instance_id IN (...) computed from replicaStates (no time window
needed). Compare folded into Config as a view-mode toggle. Two-phase
rollout (backend ships first to close the audit gap immediately).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Maven: enable useIncrementalCompilation; Surefire forkCount=1C +
reuseForks=true so unit-test JVMs are reused per CPU core instead of
spawning per class (205 tests pass under the new strategy).
- Testcontainers: opt-in reuse via .withReuse(true) on Postgres +
ClickHouse base; per-developer enable via ~/.testcontainers.properties.
- UI: drop redundant `tsc --noEmit` from `npm run build` (Vite already
type-checks); split into a dedicated `npm run typecheck` script.
- CI: cache ~/.npm and ui/node_modules/.vite alongside Maven; npm ci with
--prefer-offline --no-audit --fund=false; paths-ignore for docs-only,
.planning/ and .claude/ changes so doc-only pushes skip the pipeline.
- Docs: CLAUDE.md + .claude/rules/cicd.md updated with the new build
knobs and the Testcontainers reuse opt-in.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Backend: rename deleteTerminalByAppAndEnvironment → deleteFailedByAppAndEnvironment.
STOPPED rows were being wiped on every redeploy, so Checkpoints was always empty.
Now only FAILED rows are pruned; STOPPED deployments are retained as restorable
checkpoints (they still carry deployed_config_snapshot from their RUNNING window).
- UI filter: any deployment with a snapshot is a checkpoint (was RUNNING|DEGRADED only,
which excluded the main case — the previous blue/green deployment now in STOPPED).
- UI placement: Checkpoints disclosure now renders inside IdentitySection, matching
the design spec.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Refresh the three rules files to match the new executor behavior:
- docker-orchestration.md: rewrite DeploymentExecutor Details with
container naming scheme ({...}-{replica}-{generation}), strategy
dispatch (blue-green vs rolling), and the new DEGRADED semantics
(post-deploy only). Update TraefikLabelBuilder + ContainerLogForwarder
bullets for the generation suffix + new cameleer.generation label.
- app-classes.md: DeploymentExecutor + TraefikLabelBuilder bullets
mirror the same.
- core-classes.md: add DeploymentStrategy enum; note DEGRADED is now
post-deploy-only.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Resources tab: add a hint under the Deploy Strategy dropdown that
explains the blue-green vs rolling trade-off (resource peak, failure
semantics), switching text based on the current selection.
StatusCard: show the active deployment's strategy inline in the info
grid so users can tell at a glance which path was taken for a given
deployment.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Four ITs covering strategy behavior:
- BlueGreenStrategyIT#blueGreen_allHealthy_stopsOldAfterNew:
old is stopped only after all new replicas are healthy.
- BlueGreenStrategyIT#blueGreen_partialHealthy_preservesOldAndMarksFailed:
strict all-healthy — one starting replica aborts the deploy and
leaves the previous deployment RUNNING untouched.
- RollingStrategyIT#rolling_allHealthy_replacesOneByOne:
InOrder on stopContainer confirms old-0 stops before old-1 (the
interleaving that distinguishes rolling from blue-green).
- RollingStrategyIT#rolling_failsMidRollout_preservesRemainingOld:
mid-rollout health failure stops only the in-flight new containers
and the already-replaced old-0; old-1 stays untouched.
Shortens healthchecktimeout to 2s via @TestPropertySource so failure
paths complete in ~25s instead of ~60s.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace the Phase 3 stub with a working rolling implementation.
Flow:
- Capture previous deployment's per-index container ids up front.
- For i = 0..replicas-1:
- Start new[i] (gen-suffixed name, coexists with old[i]).
- Wait for new[i] healthy (new waitForOneHealthy helper).
- On success: stop old[i] if present, continue.
- On failure: stop in-flight new[0..i], leave un-replaced old[i+1..N]
running, mark FAILED. Already-replaced old replicas are not
restored — rolling is not reversible; user redeploys to recover.
- After the loop: sweep any leftover old replicas (when replica count
shrank) and mark the old deployment STOPPED.
Resource peak: replicas + 1.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Phase 3 of deployment-strategies plan. Refactor executeAsync to
dispatch on DeploymentStrategy.fromWire(config.deploymentStrategy()).
Blue-green (default):
- Start all N new replicas (gen-suffixed names coexist with old).
- Wait for ALL healthy (strict — partial-healthy = FAILED, preserves
previous deployment untouched).
- Only then find + stop the previous deployment.
- Final status is always RUNNING; DEGRADED is now reserved for
post-deploy replica crashes (set by DockerEventMonitor).
Rolling: stub — throws UnsupportedOperationException for now, gets
its real implementation in Phase 4.
Refactor details:
- Extract DeployCtx record to carry 13 per-deploy values around.
- Extract startReplica(ctx, i, stateOut) — shared by both strategy paths.
- Extract persistSnapshotAndMarkRunning(ctx, primaryCid) — shared finalizer.
- Rename waitForAnyHealthy → waitForAllHealthy (the name was misleading;
the method already waited for all, just returned partial on timeout).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Append an 8-char generation id (first 8 chars of deployment UUID) to:
- container name: {tenant}-{env}-{app}-{replica}-{gen}
- CAMELEER_AGENT_INSTANCEID (so old+new agents are distinct in the registry)
- Traefik cameleer.instance-id label
And emit a new standalone cameleer.generation label so dashboards
(Prometheus/Grafana) can pin deploy boundaries without regex on
instance-id.
Strategy branching comes next — this commit is foundation only; the
interim destroy-then-start flow still runs regardless of strategy.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Typed enum (BLUE_GREEN, ROLLING) with fromWire/toWire kebab-case
translation. fromWire falls back to BLUE_GREEN for unknown or null
input so the executor dispatch site never null-checks and no
misconfigured container-config can throw at runtime.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
7-phase plan to replace the interim destroy-then-start flow (f8dccaae)
with a strategy-aware executor. Adds gen-suffixed container names so
old + new replicas can coexist, plus a cameleer.generation label for
Prometheus/Grafana deploy-boundary annotations.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previous ring approach was too subtle against most env colors. Replace
the DS Avatar with a purpose-built circle rendered in the environment's
chosen color, showing 1–2 letter initials in white. Fills the full
circle so the color reads at a glance from across the list.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Container names are deterministic: {tenant}-{envSlug}-{appSlug}-{replica}.
The prior code did the stop-existing step at SWAP_TRAFFIC, *after*
START_REPLICAS had already tried to create containers with the same
names — so a redeploy against a RUNNING app consistently failed with
Docker 409 "container name already in use".
Move the stop-existing block to run right after CREATE_NETWORK and
before START_REPLICAS. SWAP_TRAFFIC becomes a label-only marker (traffic
is swapped implicitly by Traefik labels once new replicas are healthy).
Also: add `findActiveByAppIdAndEnvironmentIdExcluding` so the SQL
excludes the current deployment by id — previously the Java-side
`!id.equals(me)` guard failed because the newly-inserted row has
status=STARTING (DB default) and ORDER BY created_at DESC LIMIT 1
picked the new row, hiding the actual previous deployment.
Trade-off: this is destroy-then-start rather than true blue/green —
brief downtime during the swap. Matches the pre-unified-page behavior
and is what users reasonably expect. True blue/green would require
per-deployment container names.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
1. Add a 'Pending deploy' Badge next to the app name when there are
local edits or the saved state differs from the last deploy. Makes
the undeployed-changes state visible even when the user isn't looking
at the tab asterisks.
2. Move Start/Stop buttons from StatusCard into the page header, next
to Delete. Runs off the latest deployment's status — Stop when
RUNNING/STARTING/DEGRADED, Start (triggers a redeploy of the last
version) when STOPPED. DeploymentTab and StatusCard shed their
onStop/onStart props.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
DeploymentExecutor already persists errorMessage on FAILED transitions
but the UI never rendered it — users saw "FAILED" with no explanation.
Add a bordered error block above the action row when a deployment is
FAILED, preserving whitespace and wrapping long Docker error bodies
(e.g. 409 conflict JSON).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wrap Avatar in a span with box-shadow outline in the environment's
chosen color (slate/red/amber/green/teal/blue/purple/pink). Applied to
both the list row and the detail header. Keeps the Avatar's name-hash
interior so initials remain distinguishable; the ring just signals
which env you're looking at at a glance.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace empty-body ResponseEntity.status(BAD_REQUEST).build() with
ResponseStatusException so Spring returns the usual error body shape
with a descriptive reason string, matching the idiom used by
UserAdminController, AppSettingsController, ThresholdAdminController.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previously `dirtyState?.dirty ?? true` caused a stale `Redeploy` label
to flash briefly while the first fetch was in flight. Gate the default
on isLoading so the button starts as `Save (disabled)` until the
endpoint resolves — spurious Redeploy clicks were harmless but the
loading-state UX was wrong.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Drop the Number.isInteger normalization hack in useDeploymentPageState
that mapped 1.0 → "1.0" but broke for values like 1.10 (which round-trip
to 1.1). Instead, useFormDirty now parseFloats samplingRate on both sides
before comparing, so "1", "1.0", and "1.00" all compare equal regardless
of how the backend serializes the number.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds @Parameter description so the generated OpenAPI spec / Swagger UI
explains what 'staged' vs 'live' means instead of just surfacing the
bare param name. Follow-up: run `cd ui && npm run generate-api:live`
against a live backend to refresh openapi.json + schema.d.ts.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
On long deployment histories the StartupLogPanel would render off-screen
when the user clicked a row. Ref + useEffect scrolls the panel into view
with block:'nearest' so expanding a row that's already in view doesn't
cause a disorienting jump.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Existing IT only exercises the startContainer-throws path, where the
exception bypasses the entire try block. Add a test where startContainer
succeeds but getContainerStatus never returns healthy — this covers the
early-exit at the HEALTH_CHECK stage, which is the common real-world
failure shape and closest to the snapshot-write point.
Shortens healthchecktimeout to 2s via @TestPropertySource so the test
completes in a few seconds.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
handleRestore previously returned silently when deployedConfigSnapshot
was null, leaving the user wondering why their click did nothing. Show
a warning toast explaining that the checkpoint predates snapshotting.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Snapshot is written by DeploymentExecutor before the RUNNING/DEGRADED
split, so DEGRADED rows already carry a deployed_config_snapshot. Treat
them as checkpoints — partial-healthy deploys still produced a working
config worth restoring. Aligns repo query with UI filter.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Handoff summarises the unified deployment page implementation (spec,
plan, 43 commits, opened Gitea issues #147 and #148), open gaps, and
recommended kickoff for the next session.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Bug 1: default serverDirtyAgainstDeploy to true (not false) while
dirtyState query is loading — prevents the button showing 'Save'
instead of 'Redeploy' on apps with no successful deployment yet.
- Bug 2: normalize samplingRate from server as '<n>.0' when the value
is a whole-number float so serverState matches form after save,
eliminating spurious dirty detection that kept Save enabled.
- Bug 3: add success toast after handleSave completes.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Issue 1: add List<String> sensitiveKeys as 4th field to DeploymentConfigSnapshot; populate
from agentConfig.getSensitiveKeys() in DeploymentExecutor; handleRestore hydrates from
snap.sensitiveKeys directly; Deployment type in apps.ts gains sensitiveKeys field
- Issue 2: after createApp succeeds, refetchQueries(['apps', envSlug]) before navigate so the
new app is in cache before the router renders the deployed view (eliminates transient Save-
disabled flash)
- Issue 3: useDeploymentPageState useEffect now uses prevServerStateRef to detect local edits;
background refetches only overwrite form when no local changes are present
- Issue 5: handleRedeploy invalidates dirty-state + versions queries after createDeployment
resolves; handleSave invalidates dirty-state after staged save
- Issue 10: DirtyStateCalculator strips volatile agentConfig keys (version, updatedAt, updatedBy,
environment, application) before JSON comparison via scrubAgentConfig(); adds
versionBumpDoesNotMarkDirty test
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
AppsTab.tsx shrunk from 1387 to 109 lines — router now owns /apps/new
and /apps/:slug via AppDeploymentPage; list-only file retained.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add deployedConfigSnapshot field to Deployment interface (mirrors server shape)
- Remove the Task 10.3 cast in handleRestore now that the type has the field
- New useUnsavedChangesBlocker hook (react-router useBlocker, v7.13.1)
- Wire AlertDialog into AppDeploymentPage for in-app navigation guard
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
DeploymentTab composes StatusCard, DeploymentProgress, StartupLogPanel,
and HistoryDisclosure for the latest deployment. StartupLogPanel gains an
optional className prop, drops the fixed maxHeight, and its .panel rule
uses flex-column + min-height:0 so a parent can drive its height.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Collapsible deployment history table (sorted newest-first) with
click-to-expand StartupLogPanel for any historical deployment row.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Status badge, replica count, URL, JAR/checksum grid, and stop/start
actions for the latest deployment. CSS added to AppDeploymentPage.module.css.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Ports the ConfigSubTab traces/taps and route recording content into
standalone tab components. Each write goes straight to live agents via
useUpdateApplicationConfig (apply='live'). A local draft state prevents
stale reads during the async flush. LiveBanner is rendered at the top of
both tabs to communicate the live-apply semantics.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a warning banner that communicates live-apply semantics (changes
bypass the Save/Redeploy cycle). Uses --warning-bg / --warning-border
DS tokens. CSS class .liveBanner added to AppDeploymentPage.module.css.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pure presentational tab receiving SensitiveKeysFormState via value/onChange.
Calls useSensitiveKeys() internally to show global baseline (readonly).
Local useState for the new-key input buffer. Reuses skStyles from
SensitiveKeysPage.module.css for consistent pill/badge layout.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pure presentational tab receiving VariablesFormState via value/onChange.
Rows use the new .envVarsList / .envVarRow CSS grid (1fr 2fr auto).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pure presentational tab receiving ResourcesFormState via value/onChange.
Local useState buffers for newPort/newNetwork keep the "add next item"
inputs isolated from form state. isProd prop gates the memory-reserve field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pure presentational tab receiving MonitoringFormState via value/onChange.
Also adds shared config-tab styles to AppDeploymentPage.module.css
(configInline, toggleEnabled/Disabled, portPills, inputSizes, envVarsList/Row).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Picks up GET dirty-state, PUT config ?apply=staged|live, and
deployedConfigSnapshot on Deployment for the deployment config-diff UI.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Wires DirtyStateCalculator behind an HTTP endpoint on AppController.
Adds findLatestSuccessfulByAppAndEnv to PostgresDeploymentRepository,
registers DirtyStateCalculator as a Spring bean (with ObjectMapper for
JavaTimeModule support), and covers all three scenarios with IT.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- compareJson now recurses when both nodes are ObjectNode, so nested maps
(tracedProcessors, routeRecording, routeSamplingRates) produce deep paths
like agentConfig.tracedProcessors.proc-1 instead of a blob diff
- Extract nodeToString helper: value nodes use asText() (strips JSON quotes),
null becomes "(none)", arrays/objects get compact JSON
- Apply nodeToString in both diff-emission paths (top-level mismatch + leaf)
- Add three new tests: nullAgentConfigInSnapshot, nestedAgentField_reportsDeepPath,
stringField_differenceValueIsUnquoted (8 tests total, all pass)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Pure-logic dirty-state detection: compares desired JAR + agent config + container
config against the DeploymentConfigSnapshot from the last successful deployment.
Returns a structured DirtyStateResult with per-field differences. 5 unit tests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add @DirtiesContext(AFTER_CLASS) so the SpyBean-forked context is torn
down after the 6 tests finish, preventing permanent cache pollution
- Replace single-row queryForObject with queryForList + hasSize(1) in both
audit tests so spurious extra rows will fail explicitly
- Assert auditCount == 0 in the 400 test to lock in the no-audit-on-bad-input invariant
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace the misleading putConfig_staged_auditActionIsStagedAppConfig test
(which only checked pushResult.total == 0, a duplicate of _savesButDoesNotPush)
with two real audit-log assertions: one verifying "stage_app_config" is written
for apply=staged and a new companion test verifying "update_app_config" for the
live path. Uses jdbcTemplate to query audit_log directly (Option B).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When apply=staged, saves to DB only — no CONFIG_UPDATE dispatched to agents.
When apply=live (default, back-compat), preserves today's immediate-push behavior.
Unknown apply values return 400. Audit action is stage_app_config vs update_app_config.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds the missing containerConfig assertion to snapshot_isPopulated_whenDeploymentReachesRunning
(runtimeType + appPort entries), and tightens the await predicate from .isIn(RUNNING, DEGRADED)
to .isEqualTo(RUNNING) — the mock returns a healthy container so RUNNING is deterministic.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Injects PostgresApplicationConfigRepository into DeploymentExecutor and
calls saveDeployedConfigSnapshot at the COMPLETE stage, before
markRunning. Snapshot contains jarVersionId, agentConfig (nullable),
and app.containerConfig. The FAILED catch path is left untouched so
snapshot stays null on failure. Verified by DeploymentSnapshotIT.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace manual `new PostgresDeploymentRepository(jdbcTemplate, new ObjectMapper())` with
`@Autowired PostgresDeploymentRepository repository` to use the Spring-managed bean whose
ObjectMapper has JavaTimeModule registered. Also removes the redundant isNotNull() assertion
whose work is done by the field-level assertions that follow.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Appends DeploymentConfigSnapshot deployedConfigSnapshot to the Deployment
record and adds a matching withDeployedConfigSnapshot wither. All
positional call sites (repository mapper, test fixture) updated to pass
null; Task 1.4 will wire real persistence and Task 1.5 will populate
the field on RUNNING transition.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
All other FKs to app_versions.id (e.g. Deployment.appVersionId) use UUID;
DeploymentConfigSnapshot.jarVersionId was incorrectly typed as String.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
13 phases, TDD-oriented: Flyway V3 snapshot column, staged/live config
write flag, dirty-state endpoint, regen OpenAPI, then the new React page
(Identity, Checkpoints, 7 tabs including the live-apply Traces+Taps and
Route Recording with banner), primary Save/Redeploy state machine,
router blocker, old view cleanup, rules docs, and a manual QA walkthrough.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Single page at /apps/:slug (+ /apps/new in net-new mode) replacing the
CreateAppView/AppDetailView split. Save ↔ Redeploy state machine driven
by a deployment snapshot on the deployments table, agent-config writes
gain ?apply=staged|live, Identity & Artifact always visible, new
Deployment tab carries progress + startup log, and checkpoints restore
full prior state (JAR + config) from past successful deploys.
Concurrent-edit protection deferred to #147.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Replace EnvironmentSelector "All Envs" dropdown with Button+Modal (DS Modal, forced on first-use).
- Add 8-swatch preset color picker in the Environment settings "Appearance" section; commits via useUpdateEnvironment.
- Render a 3px fixed top bar in the current env's color across every page (z-index 900, below DS modals).
- New env-colors tokens (--env-color-*, light + dark) and envColorVar() helper with slate fallback.
- Vitest coverage for button, modal, and color helpers (13 new specs).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
UpdateEnvironmentRequest gains an optional color; Environment schema
surfaces color on GET responses.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace env dropdown with button+modal pattern, remove All Envs,
add 8-swatch preset color palette per env rendered as 3px top bar.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
jmustache on the classpath (for alert notification templates) triggers
Spring Boot's MustacheAutoConfiguration, which warns about the missing
classpath:/templates/ folder we don't use. Disable its check.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fix stale `AGGREGATE` label (actual enum: `COUNT_IN_WINDOW`). Expand
EXCHANGE_MATCH section with both fire modes, PER_EXCHANGE config-surface
restrictions (0 for reNotifyMinutes/forDurationSeconds, at-least-one-sink
rule), exactly-once guarantee scope, and the first-run backlog-cap knob.
Surface the new config in application.yml with the 24h default and the
opt-out-to-0 semantics.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New property cameleer.server.alerting.perExchangeDeployBacklogCapSeconds
(default 86400 = 24h, 0 disables). On first run (no persisted cursor
or malformed), clamp cursorTs to max(rule.createdAt, now - cap) so a
long-lived PER_EXCHANGE rule doesn't scan from its creation date
forward on first post-deploy tick. Normal-advance path unaffected.
Follows up final-review I-1 on the PER_EXCHANGE exactly-once phase.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Task 6.2 housekeeping — add BatchResultApplier to the class map per
CLAUDE.md convention. Introduced in Task 2.2 as the @Transactional
wrapper for atomic per-rule batch commits (instance writes + notification
enqueues + cursor advance).
Also refreshes GitNexus index stats auto-emitted into AGENTS.md /
CLAUDE.md (8778 -> 8893 nodes, 22647 -> 23049 edges).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
End-to-end lifecycle test: 5 FAILED exchanges across 2 ticks produces
exactly 5 FIRING instances + 5 PENDING notifications. Tick 3 with no
new exchanges produces zero new instances or notifications. Ack on one
instance leaves the other four untouched.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wire up the existing POST /alerts/rules/{id}/render-preview endpoint
so rule authors can preview their Mustache-templated notification
before saving. Available in edit mode only (new rules require save
first — endpoint is id-bound). Matches the njams gap: their rules
builder ships no in-builder preview and operators compensate with
trial-and-error save/retry.
Implementation notes:
- ReviewStep gains an optional `ruleId` prop; when present, a
"Preview notification" button calls `useRenderPreview` (the
existing TanStack mutation in api/queries/alertRules.ts) and
renders title + message in a titled, read-only pane styled like
a notification card.
- Errors surface as a DS Alert (variant=error) beneath the button.
- `RuleEditorWizard` passes `ruleId={id}` through — mirrors the
existing TriggerStep / NotifyStep wiring.
- No stateless (/render-preview without id) variant exists on the
backend, so for new rules the button is simply omitted.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Shows a warning banner and disables the Save button when a rule has
neither webhooks nor targets — would have been rejected at the server
edge (Task 3.3 validator), now caught earlier in the wizard with clear
reason.
Disable reNotifyMinutes at 0 with tooltip when PER_EXCHANGE is selected
(server rejects non-zero per Task 3.3 validator). Hide forDurationSeconds
entirely for PER_EXCHANGE (not applicable to per-exchange semantics).
Values stay zeroed via Task 4.3's applyFireModeChange helper on any
mode toggle.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Prevents stale COUNT_IN_WINDOW threshold/windowSeconds from surviving
PER_EXCHANGE save (would trip the Task 3.3 server-side validator).
Also forces reNotifyMinutes=0 and forDurationSeconds=0 when switching to
PER_EXCHANGE.
Turns green: form-state.test.ts#applyFireModeChange (3 tests).
Three failing tests pinning Task 4.3's mode-toggle state hygiene:
- clears threshold+windowSeconds on COUNT_IN_WINDOW -> PER_EXCHANGE
- returns to defaults (not stale values) on PER_EXCHANGE -> COUNT_IN_WINDOW
- forces reNotifyMinutes=0 and forDurationSeconds=0 on PER_EXCHANGE
Targets a to-be-introduced pure helper `applyFireModeChange(form, newMode)`
in form-state.ts. Task 4.3 will implement the helper and wire it into
ExchangeMatchForm so the Fire-mode <Select> calls it instead of the current
raw patch({ fireMode }) that leaves stale fields.
Follows backend removal of the field (Task 3.1). Typechecker confirms
zero remaining references. The ExchangeMatchForm linger-input is
visually removed in Task 4.4.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PER_EXCHANGE rules: 400 if reNotifyMinutes != 0 or forDurationSeconds != 0.
Any rule: 400 if webhooks + targets are both empty (never notifies anyone).
Turns green: AlertRuleControllerIT#createPerExchangeRule_with*NonZero_returns400,
AlertRuleControllerIT#createAnyRule_withEmptyWebhooksAndTargets_returns400.
Three failing IT tests documenting the contract Task 3.3 will satisfy:
- createPerExchangeRule_withReNotifyMinutesNonZero_returns400
- createPerExchangeRule_withForDurationSecondsNonZero_returns400
- createAnyRule_withEmptyWebhooksAndTargets_returns400
Dead field — was enforced by compact ctor as required for PER_EXCHANGE,
but never read anywhere in the codebase. Removal tightens the API surface
and is precondition for the Task 3.3 cross-field validator.
Pre-prod; no shim / migration.
Wraps instance writes, notification enqueues, and cursor advance in one
transactional boundary per rule tick. Rollback leaves the rule replayable
on next tick. Turns the Phase 2 atomicity IT green (see AlertEvaluatorJobIT
#tickRollback_faultOnSecondNotificationInsert_leavesCursorUnchanged).
Fault-injection IT asserts that a crash mid-batch rolls back every
instance + notification write AND leaves the cursor unchanged. Fails
against current (Phase 1 only) code — turns green when Task 2.2
wraps batch processing in @Transactional.
Replace async @AfterEach ALTER...DELETE with @BeforeEach TRUNCATE TABLE
executions — matches the convention used in ClickHouseExecutionStoreIT
and peers. Env-slug isolation was already preventing cross-test pollution;
this change is about hygiene and determinism (TRUNCATE is synchronous).
Thread EvalResult.Batch.nextEvalState into releaseClaim so the composite
cursor from Task 1.5 actually lands in rule.evalState across tick boundaries.
Guards against empty-batch wipe (would regress to first-run scan).
Follow-up to Task 1.2 flagged by Task 1.5 review (I-1). Single-column
ORDER BY could drop tail rows in a same-millisecond group >50 when
paginating via the composite cursor. Appending ', execution_id <dir>'
as secondary key only when afterExecutionId is set preserves existing
behaviour for UI/stats callers.
Two failing tests documenting the contract Task 1.5 will satisfy:
- cursorMonotonicity_sameMillisecondExchanges_fireExactlyOncePerTick
- firstRun_boundedByRuleCreatedAt_notRetentionHistory
Compile may fail until Task 1.4 adds AlertRule.withEvalState wither.
Adds an optional afterExecutionId field to SearchRequest. When combined
with a non-null timeFrom, ClickHouseSearchIndex applies a strictly-after
tuple predicate (start_time > ts OR (start_time = ts AND execution_id > id))
so same-millisecond exchanges can be consumed exactly once across ticks.
When afterExecutionId is null, timeFrom keeps its existing >= semantics —
no behaviour change for any current caller.
Also adds the SearchRequest.withCursor(ts, id) wither. Threads the field
through existing withInstanceIds / withEnvironment witheres. All existing
positional call-sites (SearchController, ExchangeMatchEvaluator,
ClickHouseSearchIndexIT, ClickHouseChunkPipelineIT) pass null for the new
slot.
Task 1.2 of docs/superpowers/plans/2026-04-22-per-exchange-exactly-once.md.
The evaluator-side wiring that actually supplies the cursor is Task 1.5.
Plan for executing the tightened spec. TDD per task: RED test first,
minimal GREEN impl, commit. Phases 1-2 land the cursor + atomic batch
commit; phase 3 validates config; phase 4 fixes the UI mode-toggle
leakage + empty-targets guard + render-preview pane; phases 5-6 close
with full-lifecycle IT and regression sweep.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Correct the factual claim that the cursor advances — it is dead code:
_nextCursor is computed but never persisted by applyBatchFiring/reschedule,
so every tick re-enqueues notifications for every matching exchange in
retention. Clarify that instance-level dedup already works via the unique
index; notification-level dedup is what's broken. Reframe §2 as "make it
atomic before §1 goes live."
Add builder-UX lessons from the njams Server_4 rules editor: clear stale
fields on fireMode toggle (not just hide them); block save on empty
webhooks+targets; wire the already-existing /render-preview endpoint into
the Review step. Add Test 5 (red-first notification-bleed regression) and
Test 6 (form-state clear on mode toggle).
Park two follow-ups explicitly: sealed condition-type hierarchy (backend
lags the UI's condition-forms/* sharding) and a coalesceSeconds primitive
for Inbox-storm taming. Amend cursor-format-churn risk: benign in theory,
but first post-deploy tick against long-standing rules could scan from
rule.createdAt forward — suggests a deployBacklogCap clamp to bound the
one-time backlog flood.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Four focused correctness fixes for the "fire exactly once per FAILED
exchange" use case (alerting layer only; HTTP-level idempotency is a
separate scope):
1. Composite cursor (startTime, executionId) replaces the current
single-timestamp, inclusive cursor — prevents same-millisecond
drops and same-exchange re-selection.
2. First-run cursor initialized to rule createdAt (not null) —
prevents the current unbounded historical-retention scan on first
tick of a new rule.
3. Transactional coupling of instance writes + notification enqueue +
cursor advance — eliminates partial-progress failure modes on crash
or rollback.
4. Config hygiene: reNotifyMinutes forced to 0, forDurationSeconds
rejected, perExchangeLingerSeconds removed entirely (was validated
as required but never read) — the rule shape stops admitting
nonsensical PER_EXCHANGE combinations.
Alert stays FIRING until human ack/resolve (no auto-resolve); webhook
fires exactly once per AlertInstance; Inbox never sees duplicates.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- AGENTS.md / CLAUDE.md: GitNexus stat block re-rendered by the analyze
hook after the last indexing run (8778 symbols / 22647 relationships).
- Remove checked-in ci-log.txt and ci-log2.txt — leftover debug output
from an earlier CI troubleshooting session, not referenced anywhere.
Also deleted untracked ui/playwright.config.js and ui/vitest.config.js
from the working tree — those are stray compiled-to-JS artifacts of the
tracked .ts config sources, not intended to be committed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
closeBrackets auto-inserts `}}` when the user types `{{`, so the buffer
already reads `{{<prefix>}}` before a completion is accepted. The apply
callback was unconditionally appending another `}}`, producing
`{{path}}}}` (valid Mustache but obviously wrong).
Fix: peek at the two characters immediately after the completion range
and, when they're `}}`, extend the replacement range by two so the
existing closing braces are overwritten rather than left in place.
Added a regression test that drives `apply` through a real EditorView
for both the bare-prefix (no trailing `}}`) and auto-closed
(`{{prefix}}`) scenarios.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Before: the Silences page rendered an always-visible 4-field form strip
above the list, taking room even when the environment had zero silences.
Inconsistent with Rules, which puts a "New rule" action in the page
header and reserves the content area for either the list or an empty
state.
After: header mirrors Rules — title + subtitle on the left, a "New
silence" primary button on the right. The create form moved into a
Modal opened by that button (and by the empty-state's "Create silence"
action). `?ruleId=` deep links still work: the param is read on mount,
prefills the Rule ID field, and auto-opens the modal — preserving the
InboxPage "Silence rule… → Custom…" flow.
Dropped: unused `sectionStyles` import.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Records the three fix commits + two prod-code cleanup commits, with
one-paragraph summaries for each cluster and pointers to the diagnosis
doc for SSE.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three distinct root causes, all reproducible when the classes run
solo — not order-dependent as the triage report suggested. Full
diagnosis in .planning/sse-flakiness-diagnosis.md.
1. AgentSseController.events auto-heal was over-permissive: any valid
JWT allowed registering an arbitrary path-id, a spoofing vector.
Surface symptom was the parked sseConnect_unknownAgent_returns404
test hanging on a 200-with-empty-stream instead of getting 404.
Fix: auto-heal requires JWT subject == path id.
2. SseConnectionManager.pingAll read ${agent-registry.ping-interval-ms}
(unprefixed). AgentRegistryConfig binds cameleer.server.agentregistry.*
— same family of bug as the MetricsFlushScheduler fix in a6944911.
Fix: corrected placeholder prefix.
3. Spring's SseEmitter doesn't flush response headers until the first
emitter.send(); clients on BodyHandlers.ofInputStream blocked on
the first body byte, making awaitConnection(5s) unreliable under a
15s ping cadence. Fix: send an initial ": connected" comment on
connect() so headers hit the wire immediately.
Verified: 9/9 SSE tests green across AgentSseControllerIT + SseSigningIT.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
No callers after the legacy PG ingestion path was retired in 0f635576.
core-classes.md updated to drop the leftover note.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After the ExecutionController removal (0f635576), SearchIndexer
subscribed to ExecutionUpdatedEvent but nothing publishes that event.
Every SearchIndexerStats metric returned always-zero, and the admin
/api/v1/admin/clickhouse/pipeline endpoint that surfaced those stats
carried no signal.
Backend removed:
- core: SearchIndexer, SearchIndexerStats, ExecutionUpdatedEvent
- app: IndexerPipelineResponse DTO, /pipeline endpoint on
ClickHouseAdminController (field + ctor param)
- StorageBeanConfig.searchIndexer bean
UI removed:
- IndexerPipeline type + useIndexerPipeline hook in
api/queries/admin/clickhouse.ts
- Indexer Pipeline card in ClickHouseAdminPage.tsx (plus ProgressBar
import and pipeline* CSS classes)
OpenAPI schema.d.ts + openapi.json regenerated (stale /pipeline path
and IndexerPipelineResponse schema removed).
SearchIndex interface + ClickHouseSearchIndex impl kept — those are
live and used by SearchService + ExchangeMatchEvaluator.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The @Scheduled placeholder read ${ingestion.flush-interval-ms:1000}
(unprefixed) but IngestionConfig binds cameleer.server.ingestion.* —
YAML tuning of the metrics flush interval was silently ignored and the
scheduler fell back to the 1s default in every environment.
Corrected to ${cameleer.server.ingestion.flush-interval-ms:1000}.
(The initial attempt to bind via SpEL #{@ingestionConfig.flushIntervalMs}
failed because beans registered via @EnableConfigurationProperties use a
compound bean name "<prefix>-<FQN>", not the simple camelCase form. The
property-placeholder path is sufficient — IngestionConfig still owns
the Java-side default.)
BackpressureIT: drops the obsolete workaround property
`ingestion.flush-interval-ms=60000`; the single prefixed override now
controls both buffer config and flush cadence.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two-layer fix for the TZ drift that caused stats reads to miss every row
when the JVM default TZ and CH session TZ disagreed:
- Insert side: ClickHouse JDBC 0.9.7 formats java.sql.Timestamp via
Timestamp.toString(), which uses JVM default TZ. A CEST JVM shipping
to a UTC CH server stored Unix timestamps off by the TZ offset (the
triage report's original symptom). Pinned JVM default to UTC in
CameleerServerApplication.main() — standard practice for observability
servers that push to time-series stores.
- Read side: stats_1m_* tables now declare bucket as DateTime('UTC'),
MV SELECTs wrap toStartOfMinute(start_time) in toDateTime(..., 'UTC')
so projections match column type, and ClickHouseStatsStore.lit(Instant)
emits toDateTime('...', 'UTC') rather than a bare literal — defence
in depth against future refactors.
Test class pins its own JVM TZ (the store IT builds its own
HikariDataSource, bypassing the main() path). Debug scaffolding from
the triage investigation removed.
Greenfield CH — no migration needed.
Verified: 14/14 ClickHouseStatsStoreIT green, plus 84/84 across all
ClickHouse IT classes (no regression from the JVM TZ default change).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Task-by-task plan for the 2026-04-21-it-triage-followups-design spec.
Autonomous execution variant — SSE diagnose-then-fix branches to either
apply-fix or park-with-@Disabled based on diagnosis confidence, since
this runs unattended overnight.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Design for closing the 12 parked IT failures (ClickHouseStatsStoreIT
timezone, SSE flakiness in AgentSseControllerIT/SseSigningIT) plus two
production-code side notes the ExecutionController removal surfaced:
- ClickHouseStatsStore timezone fix — column-level DateTime('UTC') on
bucket, greenfield CH
- SSE flakiness — diagnose-then-fix with user checkpoint between phases
- MetricsFlushScheduler property-key fix — bind via SpEL, single source
of truth in IngestionConfig
- Dead-code cleanup — SearchIndexer.onExecutionUpdated listener +
unused TaggedExecution record
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ExecutionController was @ConditionalOnMissingBean(ChunkAccumulator.class),
and ChunkAccumulator is registered unconditionally — the legacy controller
never bound in any profile. Even if it had, IngestionService.ingestExecution
called executionStore.upsert(), and the only ExecutionStore impl
(ClickHouseExecutionStore) threw UnsupportedOperationException from upsert
and upsertProcessors. The entire RouteExecution → upsert path was dead code
carrying four transitive dependencies (RouteExecution import, eventPublisher
wiring, body-size-limit config, searchIndexer::onExecutionUpdated hook).
Removed:
- cameleer-server-app/.../controller/ExecutionController.java (whole file)
- ExecutionStore.upsert + upsertProcessors (interface methods)
- ClickHouseExecutionStore.upsert + upsertProcessors (thrower overrides)
- IngestionService.ingestExecution + toExecutionRecord + flattenProcessors
+ hasAnyTraceData + truncateBody + toJson/toJsonObject helpers
- IngestionService constructor now takes (DiagramStore, WriteBuffer<Metrics>);
dropped ExecutionStore + Consumer<ExecutionUpdatedEvent> + bodySizeLimit
- StorageBeanConfig.ingestionService(...) simplified accordingly
Untouched because still in use:
- ExecutionRecord / ProcessorRecord records (findById / findProcessors /
SearchIndexer / DetailController)
- SearchIndexer (its onExecutionUpdated never fires now since no-one
publishes ExecutionUpdatedEvent, but SearchIndexerStats is still
referenced by ClickHouseAdminController — separate cleanup)
- TaggedExecution record has no remaining callers after this change —
flagged in core-classes.md as a leftover; separate cleanup.
Rule docs updated:
- .claude/rules/app-classes.md: retired ExecutionController bullet, fixed
stale URL for ChunkIngestionController (it owns /api/v1/data/executions,
not /api/v1/ingestion/chunk/executions).
- .claude/rules/core-classes.md: IngestionService surface + note the dead
TaggedExecution.
Full IT suite post-removal: 560 tests run, 11 F + 1 E — same 12 failures
in the same 3 previously-parked classes (AgentSseControllerIT / SseSigningIT
SSE-timing + ClickHouseStatsStoreIT timezone bug). No regression.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
13 commits landed on local main; the three remaining parked clusters
each need a specific intent call before the next pass can proceed:
- ClickHouseStatsStoreIT (8 failures) — timezone bug in
ClickHouseStatsStore.lit(Instant); needs a store-side fix, not a
test-side one.
- AgentSseControllerIT + SseSigningIT (4 failures) — SSE connection
timing; looks order-dependent, not spec drift.
Also flagged two side issues worth a follow-up PR:
- ExecutionController legacy path is dead code.
- MetricsFlushScheduler.@Scheduled reads the wrong property key and
silently ignores the configured flush interval in production.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The pushToAgents fan-out iterates every distinct (app, env) slice in
the shared agent registry. In isolated runs that's 0, but with Spring
context reuse across IT classes we always see non-zero here. Assert
the response has a pushResult.total field (shape) rather than exact 0.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- ForwardCompatIT: send a valid ExecutionChunk envelope with extra
unknown fields instead of a bare {futureField}. Was being parsed into
an empty/degenerate chunk and rejected with 400.
- ProtocolVersionIT.requestWithCorrectProtocolVersionPassesInterceptor:
same shape fix — minimal valid chunk so the controller's 400 is not
an ambiguous signal for interceptor-passthrough.
- BackpressureIT:
* TestPropertySource keys were "ingestion.*" but IngestionConfig is
bound under "cameleer.server.ingestion.*" — overrides were ignored
and the buffer stayed at its default 50_000, so the 503 overflow
branch was unreachable. Corrected the keys.
* MetricsFlushScheduler's @Scheduled uses a *different* key again
("ingestion.flush-interval-ms"), so we override that separately to
stop the default 1s flush from draining the buffer mid-test.
* executionIngestion_isSynchronous_returnsAccepted now uses the
chunked envelope format.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ClickHouseChunkPipelineIT.setUp was loading /clickhouse/V2__executions.sql
and /clickhouse/V3__processor_executions.sql — resource paths that no
longer exist after 90083f88 collapsed the V1..V18 ClickHouse schema into
init.sql. Swapped for ClickHouseTestHelper.executeInitSql(jdbc).
ClickHouseExecutionReadIT.detailService_buildTree_withIterations was
asserting getLoopIndex() on children of a split, but DetailService's
seq-based buildTree path (buildTreeBySeq) maps FlatProcessorRecord.iteration
into ProcessorNode.iteration — not loopIndex. The loopIndex path is only
populated by buildTreeByProcessorId (the legacy ID-only fallback). Switched
the assertion to getIteration() to match the seq-driven reconstruction.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Both tests extend AbstractPostgresIT and inherit the Postgres jdbcTemplate,
which they were using to query ClickHouse-resident tables (executions,
processor_executions, route_diagrams). Now:
- DiagramLinkingIT reads diagramContentHash off the execution-detail REST
response (and tolerates JSON null by normalising to empty string, which
matches how the ingestion service stamps un-linked executions).
- IngestionSchemaIT asserts the reconstructed processor tree through the
execution-detail endpoint (covers both flattening on write and
buildTree on read) and reads processor bodies via the processor-snapshot
endpoint rather than raw processor_executions rows.
Both tests now use the ExecutionChunk envelope on POST /data/executions.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Largest Cluster B test: seeded 10 executions via the legacy RouteExecution
shape which ChunkIngestionController silently degenerates to empty chunks,
then verified via a Postgres SELECT against a ClickHouse table. Both
failure modes addressed:
- All 10 seed payloads are now ExecutionChunk envelopes (chunkSeq=0,
final=true, flat processors[]).
- Pipeline visibility probe is the env-scoped search REST endpoint
(polling for the last corr-page-10 row).
- searchGet() helper was using the AGENT token; env-scoped read
endpoints require VIEWER+, so it now uses viewerJwt (matches what
searchPost already did).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
DiagramControllerIT.postDiagram_dataAppearsAfterFlush now verifies via
GET /api/v1/environments/{env}/apps/{app}/routes/{route}/diagram instead
of a PG SELECT against the ClickHouse route_diagrams table.
DiagramRenderControllerIT seeds both a diagram and an execution on the
same route, then reads the stamped diagramContentHash off the execution-
detail REST response to drive the flat /api/v1/diagrams/{hash}/render
tests. The env-scoped endpoint only serves JSON, so SVG tests still hit
the content-hash endpoint — but the hash comes from REST now, not SQL.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Same pattern as DetailControllerIT:
- ExecutionControllerIT: all four tests now post ExecutionChunk envelopes
(chunkSeq=0, final=true) carrying instanceId/applicationId. Flush
visibility check pivoted from PG SELECT → env-scoped search REST.
- MetricsControllerIT: postMetrics_dataAppearsAfterFlush now stamps
collectedAt at now() and verifies through GET /environments/{env}/
agents/{id}/metrics with the default 1h lookback, looking for a
non-zero bucket on the metric name.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
POST /api/v1/data/executions is owned by ChunkIngestionController (the
legacy ExecutionController path is @ConditionalOnMissingBean(ChunkAccumulator)
and never binds). The old RouteExecution-shaped seed was silently parsed
as an empty ExecutionChunk and nothing landed in ClickHouse.
Rewrote the seed as a single final ExecutionChunk with chunkSeq=0 /
final=true and a flat processors[] carrying seq + parentSeq to preserve
the 3-level tree (DetailService.buildTree reconstructs the nested shape
for the API response). Execution-id lookup now goes through the search
REST API filtered by correlationId, per the no-raw-SQL preference.
Template for the other Cluster B ITs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Records the 5 commits landed this session (65 → 44 failures), the 3
accepted remaining clusters (Cluster B ingestion-payload drift, SSE
timing, small Cluster E tail), and the open questions that require
spec intent before the next pass can proceed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Both Testcontainers Postgres ITs were asserting exact counts on rows that
other classes in the shared context had already written.
- FlywayMigrationIT: treat the non-seed tables (users, server_config,
audit_log, application_config, app_settings) as "must exist; COUNT must
return a non-negative integer" rather than expecting exactly 0. The
seeded tables (roles=4, groups=1) still assert exact V1 baseline.
- ConfigEnvIsolationIT.findByEnvironment_excludesOtherEnvs: use unique
prefixed app slugs and switch containsExactlyInAnyOrder to contains +
doesNotContain, so the cross-env filter is still verified without
coupling to other tests' inserts.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The agent list moved from /api/v1/agents to /api/v1/environments/{envSlug}/agents;
the 'valid JWT returns 200' test was hitting the retired flat path and
getting 404. The other 'without JWT' cases still pass because Spring
Security rejects them at the filter chain before URL routing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Registration now requires environmentId in the body (400 if missing), so
the stale register bodies were failing every downstream test that relied
on a registered agent. Affected helpers in:
- BootstrapTokenIT (static constant + inline body)
- JwtRefreshIT (registerAndGetTokens)
- RegistrationSecurityIT (registerAgent)
- SseSigningIT (registerAgentWithAuth)
- AgentSseControllerIT (registerAgent helper)
Also in JwtRefreshIT / RegistrationSecurityIT, the "access token can reach
a protected endpoint" tests were hitting env-scoped read endpoints that
now require VIEWER+. Redirected both to the AGENT-role heartbeat endpoint
— it proves the token is accepted by the security filter without being
coupled to RBAC rules for reader endpoints.
JwtRefreshIT.refreshWithValidToken also dropped an isNotEqualTo assertion
that assumed sub-second iat uniqueness — HMAC JWTs with second-precision
claims are byte-identical when minted for the same subject within the
same second, so the old assertion was flaky by design.
SseSigningIT / AgentSseControllerIT still have SSE-connection timing
failures unrelated to registration — parked separately.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two drifts corrected:
- registerAgent helper missing required environmentId (spec: 400 if absent).
- sendGroupCommand is now synchronous request-reply: returns 200 with an
aggregated CommandGroupResponse {success,total,responded,responses,timedOut}
— no longer 202 with {targetCount,commandIds}. Updated assertions and name.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Four drifts against the current server contract, all now corrected:
- Registration body missing required environmentId (spec: 400 if absent).
- Agent list moved to env-scoped /api/v1/environments/{envSlug}/agents;
flat /api/v1/agents no longer exists.
- heartbeatUnknownAgent now auto-heals via JWT env claim (fb54f9cb);
the 404 branch is only reachable without a JWT, which the security
filter rejects before the controller sees the request.
- sseEndpoint is an absolute URL (ServletUriComponentsBuilder.fromCurrentContextPath),
so assert endsWith the path rather than equals-to-relative.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Reproduction: pause a container long enough to cross both the stale
and dead thresholds, then unpause. The agent resumes sending heartbeats
but the server keeps it shown as DEAD. Only a full container restart
(which re-registers) fixes it.
Root cause: AgentRegistryService.heartbeat() only revived STALE → LIVE.
A DEAD agent's heartbeat updated lastHeartbeat but left state unchanged.
checkLifecycle() never downgrades DEAD either (no-op in that branch),
so the agent was permanently stuck in DEAD until a register() call.
Fix: extend the revival branch to also cover DEAD. Same process; a
heartbeat is proof of liveness regardless of the previous state.
Also: AgentLifecycleMonitor.mapTransitionEvent() now emits RECOVERED
for DEAD → LIVE, mirroring its behavior for STALE → LIVE, so the
lifecycle timeline captures the transition.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The project is still greenfield (no production deployment) so this is
the last safe moment to flatten the migration archaeology before the
checksum history starts mattering for real.
Schema changes
- 18 migration files (531 lines) → one V1__init.sql (~380 lines)
declaring the final end-state: RBAC + claim mappings + runtime
management + config + audit + outbound + alerting, plus seed data
(system roles, Admins group, default environment).
- Drops the data-repair statements from V14 (firemode backfill),
V16 (subjectFingerprint migration), V17 (ACKNOWLEDGED → FIRING
coercion) — they were no-ops on any DB that starts at V1.
- Declares condition_kind_enum with AGENT_LIFECYCLE from the start
(was added retroactively by V18).
- Declares alert_state_enum with three values only (was five, then
swapped in V17) and alert_instances with read_at / deleted_at
columns from day one (was added by V17).
- alert_reads table never created (V12 created, V17 dropped).
- alert_instances_open_rule_uq built with the V17 predicate from
the start.
Test changes
- Replace V12MigrationIT / V17MigrationIT / V18MigrationIT with one
SchemaBootstrapIT that asserts the combined invariants: tables
present, alert_reads absent, enum value sets, alert_instances has
read_at + deleted_at, open_rule_uq exists and is unique, env-delete
cascade fires.
Verification
- pg_dump of the new V1 matches the pg_dump of V1..V18 applied in
sequence (bytewise modulo column order and Postgres-auto FK names).
- Full alerting IT suite (53 tests across 6 classes) green against
the new schema.
- The 47 pre-existing test failures on main (AgentRegistrationIT,
SearchControllerIT, ClickHouseStatsStoreIT, …) are unrelated and
fail identically without this change.
Developer impact
- Existing local DBs will fail checksum validation on boot. Wipe:
docker compose down -v (or drop the tenant_default schema).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extends the previous describeApiError rollout to the rest of the UI.
Two symptom classes covered:
- Bare e.message / err.message in toast descriptions would render
"undefined" on Spring error bodies (plain objects without a proper
Error prototype). Affected: OidcConfigPage (save/test/delete),
ClaimMappingRulesModal (save + test), AgentHealth (dismiss),
RouteControlBar (route action + replay).
- Inline {String(error)} on load-failure banners would render
"[object Object]". Affected: InboxPage, RulesListPage, SilencesPage,
OutboundConnectionsPage.
Not touched: auth-store, AppsTab, UsersTab — they already guard with
`e instanceof Error` and fall back to a static string; replacing the
fallback with describeApiError would be a behavioral change best
evaluated separately.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Backend
- V18 migration adds AGENT_LIFECYCLE to condition_kind_enum. Java
ConditionKind enum shipped with this value but no Postgres migration
extended the type, so any AGENT_LIFECYCLE rule insert failed with
"invalid input value for enum condition_kind_enum".
- ALTER TYPE ... ADD VALUE lives alone in its migration per Postgres
constraint that the new value cannot be referenced in the same tx.
- V18MigrationIT asserts the enum now contains all 7 kinds.
Frontend
- Add describeApiError(e) helper to unwrap openapi-fetch error bodies
(Spring error JSON) into readable strings. String(e) on a plain
object rendered "[object Object]" in toasts — the actual failure
reason was hidden from the user.
- Replace String(e) in all 13 toast descriptions across the alerting
and outbound-connection mutation paths.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Rewrite the "Infrastructure Setup" / "Run the Server" sections to
reflect what docker-compose.yml actually provides (full stack —
PostgreSQL + ClickHouse + server + UI — not just PostgreSQL). Adds:
- Step-by-step walkthrough for a first-run clean environment.
- Port map including the UI (8080), ClickHouse (8123/9000), PG (5432),
server (8081).
- Dev credentials baked into compose surfaced in one place.
- Lifecycle commands (stop/start/rebuild-single-service/wipe).
- Infra-only mode for backend-via-mvn / UI-via-vite iteration.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
GitNexus analyze --embeddings after the alerts-inbox-redesign branch
brought the graph to 8780 symbols / 22753 relationships (was 8527/22174
in AGENTS.md and 8603/22281 in CLAUDE.md). The stat-header drift between
AGENTS.md and CLAUDE.md is an artifact of separate reindexes — both now
in sync.
ui/tsconfig.app.tsbuildinfo was a stale tsc incremental-build cache
that shouldn't be tracked.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
UI: AlertStateChip.LABELS and .COLORS no longer include ACKNOWLEDGED
(dropped in V17). AlertStateChip.test.tsx test-cases trimmed to the
three remaining states. LayoutShell CMD-K now searches FIRING alerts
with acked=false (was state=[FIRING,ACKNOWLEDGED]).
Test: V17MigrationIT.open_rule_index_predicate_is_reworked replaced
with a structural-only assertion (index exists, indisunique). The
pg_get_indexdef pretty-printer varies across Postgres versions, so
predicate semantics are verified behaviorally in
PostgresAlertInstanceRepositoryIT (findOpenForRule_* +
save_rejectsSecondOpenInstanceForSameRuleAndExchange).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Used by InboxPage's 'Silence rule… → Custom…' flow to carry the alert's
ruleId into the silence creation form.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- STATE_ITEMS gains color dots (text-muted/error/success) to match SEVERITY_ITEMS
- onDeleteOne removes the deleted id from the selection Set so a follow-up bulk
action doesn't try to re-delete a tombstoned row
- drop stale comment block that described an alternative SilenceRulesForSelection
implementation not matching the shipped code
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces the old FIRING+ACK hardcoded inbox with the single filterable
inbox:
- Filter bar: Severity · Status (PENDING/FIRING/RESOLVED, default FIRING) ·
Hide acked (default on) · Hide read (default on).
- Row actions: Ack, Mark read, Silence rule… (quick menu), Delete
(OPERATOR+, soft delete with undo toast wired to useRestoreAlert).
- Bulk toolbar: Ack N · Mark N read · Silence rules · Delete N
(ConfirmDialog; OPERATOR+).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Used by InboxPage row + bulk actions to silence an alert's underlying
rule for a chosen preset window. 'Custom…' routes to
/alerts/silences?ruleId=<id> (T13 adds the prefill wire).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sidebar Alerts section now just: Inbox · Rules · Silences. The /alerts
redirect still lands in /alerts/inbox; /alerts/all and /alerts/history
routes are gone (no redirect — stale URLs 404 per clean-break policy).
Also updates sidebar-utils.test.ts to match the new 3-entry shape.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- useAlerts gains acked/read filter params threaded into query + queryKey
- new mutations: useBulkAckAlerts, useDeleteAlert, useBulkDeleteAlerts, useRestoreAlert
- all cache-invalidate the alerts list and unread-count on success
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New endpoints visible to the SPA: DELETE /alerts/{id}, POST
/alerts/{id}/restore, POST /alerts/bulk-delete, POST /alerts/bulk-ack.
GET /alerts gains tri-state acked / read query params. AlertDto now
includes readAt.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- new AlertInstanceRepository.filterInEnvLive(ids, env): single-query bulk ID validation
- AlertController.inEnvLiveIds now one SQL round-trip instead of N
- bulkMarkRead SQL: defense-in-depth AND deleted_at IS NULL
- bulkAck SQL already had deleted_at IS NULL guard — no change needed
- PostgresAlertInstanceRepositoryIT: add filterInEnvLive_excludes_other_env_and_soft_deleted
- V12MigrationIT: remove alert_reads assertion (table dropped by V17)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- GET /alerts gains tri-state acked + read query params
- new endpoints: DELETE /{id} (soft-delete), POST /bulk-delete, POST /bulk-ack, POST /{id}/restore
- requireLiveInstance 404s on soft-deleted rows; restore() reads the row regardless
- BulkReadRequest → BulkIdsRequest (shared body for bulk read/ack/delete)
- AlertDto gains readAt; deletedAt stays off the wire
- InAppInboxQuery.listInbox threads acked/read through to the repo (7-arg, no more null placeholders)
- SecurityConfig: new matchers for bulk-ack (VIEWER+), DELETE/bulk-delete/restore (OPERATOR+)
- AlertControllerIT: persistence assertions on /read + /bulk-read; full coverage for new endpoints
- InAppInboxQueryTest: updated to 7-arg listInbox signature
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The class under test was removed in da281933; the IT became a @Disabled
placeholder. Deleting per no-backwards-compat policy. Read mutation
coverage lives in PostgresAlertInstanceRepositoryIT going forward.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Remove SET state='ACKNOWLEDGED' from ack() and the ACKNOWLEDGED predicate
from findOpenForRule — both would error after V17. The final ack() + open-rule
semantics (idempotent guards, deleted_at) are owned by Task 5; this is just
the minimum to stop runtime SQL errors.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- AlertStateTransitionsTest: add null,null for readAt/deletedAt in openInstance helper;
replace firingWhenAcknowledgedIsNoOp with firing_with_ack_stays_firing_on_next_firing_tick;
convert ackedInstanceClearsToResolved to use FIRING+withAck; update section comment.
- PostgresAlertInstanceRepository: stub null,null for readAt/deletedAt in rowMapper
to unblock compilation (Task 4 will read the actual DB columns).
- All other alerting test files: add null,null for readAt/deletedAt to AlertInstance
ctor calls so the test source tree compiles; stub ACKNOWLEDGED JSON/state assertions
with FIRING + TODO Task 4 comments.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- AlertState: remove ACKNOWLEDGED case (V17 migration already dropped it from DB enum)
- AlertInstance: insert readAt + deletedAt Instant fields after lastNotifiedAt; add withReadAt/withDeletedAt withers; update all existing withers to pass both fields positionally
- AlertStateTransitions: add null,null for readAt/deletedAt in newInstance ctor call; collapse FIRING,ACKNOWLEDGED switch arm to just FIRING
- AlertScopeTest: update AlertState.values() assertion to 3 values; fix stale ConditionKind.hasSize(6) to 7 (JVM_METRIC was added earlier)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Collapse /alerts/inbox, /alerts/all, /alerts/history into a single
filterable inbox. Drop ACKNOWLEDGED from AlertState; add read_at and
deleted_at as orthogonal timestamp flags. Retire per-user alert_reads
tracking. Add Silence-rule and Delete row/bulk actions.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Allows alert rules to fire on agent-lifecycle events — REGISTERED,
RE_REGISTERED, DEREGISTERED, WENT_STALE, WENT_DEAD, RECOVERED — rather
than only on current state. Each matching `(agent, eventType, timestamp)`
becomes its own ackable AlertInstance, so outages on distinct agents are
independently routable.
Core:
- New `ConditionKind.AGENT_LIFECYCLE` + `AgentLifecycleCondition` record
(scope, eventTypes, withinSeconds). Compact ctor rejects empty
eventTypes and withinSeconds<1.
- Strict allowlist enum `AgentLifecycleEventType` (six entries matching
the server-emitted types in `AgentRegistrationController` and
`AgentLifecycleMonitor`). Custom agent-emitted event types tracked in
backlog issue #145.
- `AgentEventRepository.findInWindow(env, appSlug, agentId, eventTypes,
from, to, limit)` — new read path ordered `(timestamp ASC, insert_id
ASC)` used by the evaluator. Implemented on
`ClickHouseAgentEventRepository` with tenant + env filter mandatory.
App:
- `AgentLifecycleEvaluator` queries events in the last `withinSeconds`
window and returns `EvalResult.Batch` with one `Firing` per row.
Every Firing carries a canonical `_subjectFingerprint` of
`"<agentId>:<eventType>:<tsMillis>"` in context plus `agent` / `event`
subtrees for Mustache templating.
- `NotificationContextBuilder` gains an `AGENT_LIFECYCLE` branch that
exposes `{{agent.id}}`, `{{agent.app}}`, `{{event.type}}`,
`{{event.timestamp}}`, `{{event.detail}}`.
- Validation is delegated to the record compact ctor + enum at Jackson
deserialization time — matches the existing policy of keeping
controller validators focused on env-scoped / SQL-injection concerns.
Schema:
- V16 migration generalises the V15 per-exchange discriminator on
`alert_instances_open_rule_uq` to prefer `_subjectFingerprint` with a
fallback to the legacy `exchange.id` expression. Scalar kinds still
resolve to `''` and keep one-open-per-rule. Duplicate-key path in
`PostgresAlertInstanceRepository.save` is unchanged — the index is
the deduper.
UI:
- New `AgentLifecycleForm.tsx` wizard form with multi-select chips for
the six allowed event types + `withinSeconds` input. Wired into
`ConditionStep`, `form-state` (validation + defaults: WENT_DEAD,
300 s), and `enums.ts` options. Tests in `enums.test.ts` pin the
new option array.
- `alert-variables.ts` registers `{{agent.app}}`, `{{event.type}}`,
`{{event.timestamp}}`, `{{event.detail}}` leaves for the new kind,
and extends `agent.id`'s availability list to include `AGENT_LIFECYCLE`.
Tests (all passing):
- 5 new JSON-roundtrip cases on `AlertConditionJsonTest` (positive +
empty/zero/unknown-type rejection).
- 5 new evaluator unit tests on `AgentLifecycleEvaluatorTest` (empty
window, multi-agent fingerprint shape, scope forwarding, missing env).
- `NotificationContextBuilderTest` switch now covers the new kind.
- 119 alerting unit tests + 71 UI tests green.
Docs: `.claude/rules/{core,app,ui}` and CLAUDE.md migration list updated.
Inbox: replace 4 parallel outlined buttons with 2 context-aware ones.
When nothing is selected → "Acknowledge all firing" (primary) + "Mark all
read" (secondary). When rows are selected → the same slots become
"Acknowledge N" + "Mark N read" with counts inlined. Primary variant
gives the foreground action proper visual weight; secondary is the
supporting action. No more visually-identical disabled buttons cluttering
the bar.
History: drop the local DateRangePicker. The page now reads
`timeRange` from `useGlobalFilters()` so the top-bar TimeRangeDropdown
(1h / 3h / 6h / Today / 24h / 7d / custom) is the single source of
truth, consistent with every other time-scoped page in the app.
Replace the SegmentedTabs with multi-select ButtonGroup, matching the
topnavbar Completed/Warning/Failed/Running pattern. State dots use the
same palette as AlertStateChip (FIRING=error, ACKNOWLEDGED=warning,
PENDING=muted, RESOLVED=success). Default selection is the three "open"
states — Resolved is off by default and a single click surfaces closed
alerts without navigating to /history.
Backend: `GET /environments/{envSlug}/alerts` now accepts optional multi-value
`state=…` and `severity=…` query params. Filters are pushed down to
PostgresAlertInstanceRepository, which appends `AND state::text = ANY(?)` /
`AND severity::text = ANY(?)` to the inbox query (null/empty = no filter).
`AlertInstanceRepository.listForInbox` gained a 7-arg overload; the old 5-arg
form is preserved as a default delegate so existing callers (evaluator,
AlertingFullLifecycleIT, PostgresAlertInstanceRepositoryIT) compile unchanged.
`InAppInboxQuery.listInbox` also has a new filtered overload.
UI: InboxPage severity filter migrated from `SegmentedTabs` (single-select,
no color cues) to `ButtonGroup` (multi-select with severity-coloured dots),
matching the topnavbar status-filter pattern. `useAlerts` forwards the
filters as query params and cache-keys on the filter tuple so each combo
is independently cached.
Unit + hook tests updated to the new contract (5 UI tests + 8 Java unit
tests passing). OpenAPI types regenerated from the fresh local backend.
Round 4 smoke feedback on /alerts:
- Bell now has consistent 12px gap from env selector and user name
(wrap env + bell in flex container inside TopBar's environment prop)
- RuleEditorWizard constrained to max-width 840px (centered) and
upgraded the page title from SectionHeader to h2 pattern used by
the list pages
- Inbox: added select-all checkbox, severity SegmentedTabs filter
(All / Critical / Warning / Info), and bulk-ack actions
(Acknowledge selected + Acknowledge all firing) alongside the
existing mark-read actions
Surfaced during second smoke:
1. Notification bell moved — was first child of TopBar (left of
breadcrumb); now rendered inside the `environment` slot so it
sits between the env selector and the user menu, matching user
expectations.
2. Content tabs (Exchanges/Dashboard/Runtime/Deployments) hidden on
`/alerts/*` — the operational tabs don't apply there.
3. Inbox / All alerts filters now actually filter. `AlertController.list`
accepts only `limit` — `state`/`severity` query params are dropped
server-side. Move `useAlerts` to fetch once per env (limit 200) and
apply filters client-side via react-query `select`, with a stable
queryKey so filter toggles are instant and don't re-request. True
server-side filter needs a backend change (follow-up).
4. Novice-friendly labels:
- Inbox subtitle: "99 firing · 100 total" → "99 need attention ·
100 total in inbox"
- All alerts filter: Open/Firing/Acked/All →
"Currently open"/"Firing now"/"Acknowledged"/"All states"
- All alerts subtitle: "N shown" → "N matching your filter"
- History subtitle: "N resolved" → "N resolved alert(s) in range"
- Rules subtitle: "N total" → "N rule(s) configured"
- Silences subtitle: "N active" → "N active silence(s)" or
"Nothing silenced right now"
- Column headers: "State" → "Status", rules "Kind" → "Type",
rules "Targets" → "Notifies"
- Buttons: "Ack" → "Acknowledge", silence "End" → "End early"
Updated alerts.test.tsx and e2e selector to match new behavior/labels.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Visual regressions surfaced during browser smoke:
1. Page headers — `SectionHeader` renders as 12px uppercase gray (a
section divider, not a page title). Replace with proper h2 title
+ inline subtitle (`N firing · N total` etc.) and right-aligned
actions, styled from `alerts-page.module.css`.
2. Undefined `--space-*` tokens — the project (and `@cameleer/design-system`)
has never shipped `--space-sm|md|lg|xl`, even though many modules
(SensitiveKeysPage, alerts CSS, …) reference them. The fallback
to `initial` silently collapsed gaps/paddings to 0. Define the
scale in `ui/src/index.css` so every consumer picks it up.
3. List scrolling — DataTable was using default pagination, but with
no flex sizing the whole page scrolled. Add `fillHeight` and raise
`pageSize`/list `limit` to 200 so the table gets sticky header +
internal scroll + pinned pagination footer (Gmail-style). True
cursor-based infinite scroll needs a backend change (filed as
follow-up — `/alerts` only accepts `limit` today).
4. Title column clipping — `.titlePreview` used `white-space: nowrap`
+ fixed `max-width`, truncating message mid-UUID. Switch to a
2-line `-webkit-line-clamp` so full context is visible.
5. Notification bell badge invisible — `NotificationBell.module.css`
referenced undefined tokens (`--fg`, `--hover-bg`, `--bg`,
`--muted`). Map to real DS tokens (`--text-primary`, `--bg-hover`,
`#fff`, `--text-muted`). The admin user currently sees no badge
because the backend `/alerts/unread-count` returns 0 (read
receipts) — that's data, not UI.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two inline-style color refs in NotifyStep and TriggerStep were still
pointing at the undefined --muted token instead of the DS
--text-muted. Caught by the design-system-alignment verification
grep.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The Rules list Delete and Silences End-early flows now use DS
ConfirmDialog instead of native confirm(). Update selectors to
target the dialog's role=dialog + confirm button instead of
listening for the native `dialog` event.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Promote banner and prefill warnings now render as DS Alert components
(info / warning variants). Step body wraps in sectionStyles.section
for card affordance matching other forms in the app.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The feed-row component is replaced by DataTable column renderers and
the shared renderAlertExpanded content renderer. No callers remain.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces raw <table> with DataTable, inline-styled form with proper
FormField hints, and native confirm() end-early with ConfirmDialog
(warning variant). Adds DS EmptyState for no-silences case.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces raw <table> with DataTable, raw <select> promote control with
DS Dropdown, and native confirm() delete with ConfirmDialog. Adds DS
EmptyState with CTA for the no-rules case. Uses SectionHeader's
action slot instead of ad-hoc flex wrapper.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces custom feed rows with DataTable. Adds a DateRangePicker
filter (client-side) defaulting to the last 7 days. Client-side
range filter is a stopgap; a server-side range param is a future
enhancement captured in the design spec.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces 4-Button filter row with DS SegmentedTabs and custom row
rendering with DataTable. Shares expandedContent renderer and
severity-driven rowAccent with Inbox.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replaces custom feed-row layout with the shared DataTable shell used
elsewhere in the app. Adds checkbox selection + bulk "Mark selected
read" toolbar alongside the existing "Mark all read". Uses DS
EmptyState for empty lists, severity-driven rowAccent for unread
tinting, and renderAlertExpanded for row detail.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Drop the feed-row classes (.row, .rowUnread, .body, .meta, .time,
.message, .actions, .empty) — these are replaced by DS DataTable +
EmptyState in follow-up tasks. Keep layout helpers for page shell,
toolbar, filter bar, bulk-action bar, title cell, and DataTable
expanded content. All colors / spacing use DS tokens.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extracts the per-row detail block used by Inbox/All/History DataTables
so the three pages share one rendering. Consumes AlertDto fields that
are nullable in the schema; hides missing fields instead of rendering
placeholders.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Formats ISO timestamps as `Nm ago` / `Nh ago` / `Nd ago`, falling back
to an absolute locale date string for values older than 30 days. Used
by the alert DataTable Age column.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pure function mapping the 3-value AlertDto.severity enum to the 2-value
DataTable rowAccent prop. INFO maps to undefined (no tint) because the
DS DataTable rowAccent only supports error|warning.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Task-by-task TDD plan implementing the design spec. Splits the work
into 14 tasks: helper utilities (TDD), shared renderer, CSS token
migration, per-page rewrites (Inbox/All/History/Rules/Silences),
wizard banner migration, AlertRow deletion, E2E adaptation for
ConfirmDialog, and full verification pass. Each task produces an
atomic commit.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Rework all pages under /alerts to use @cameleer/design-system components
and tokens. Unified DataTable shell for Inbox/All/History with expandable
rows; DataTable + Dropdown + ConfirmDialog for Rules list; FormField grid
+ DataTable for Silences; DS Alert for wizard banners. Replaces undefined
CSS variables (--bg, --fg, --muted, --accent) with DS tokens and removes
raw <table>/<select>/confirm() usage.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
V13 added a partial unique index on alert_instances(rule_id) WHERE state
IN (PENDING,FIRING,ACKNOWLEDGED). Correct for scalar condition kinds
(ROUTE_METRIC / AGENT_STATE / DEPLOYMENT_STATE / LOG_PATTERN / JVM_METRIC
/ EXCHANGE_MATCH in COUNT_IN_WINDOW) but wrong for EXCHANGE_MATCH /
PER_EXCHANGE, which by design emits one alert_instance per matching
exchange. Under V13 every PER_EXCHANGE tick with >1 match logged
"Skipped duplicate open alert_instance for rule …" at evaluator cadence
and silently lost alert fidelity — only the first matching exchange per
tick got an AlertInstance + webhook dispatch.
V15 drops the rule_id-only constraint and recreates it with a
discriminator on context->'exchange'->>'id'. Scalar kinds emit
Map.of() as context, so their expression resolves to '' — "one open per
rule" preserved. ExchangeMatchEvaluator.evaluatePerExchange always
populates exchange.id, so per-exchange instances coexist cleanly.
Two new PostgresAlertInstanceRepositoryIT tests:
- multiple open instances for same rule + distinct exchanges all land
- second open for identical (rule, exchange) still dedups via the
DuplicateKeyException fallback in save() — defense-in-depth kept
Also fixes pre-existing PostgresAlertReadRepositoryIT brokenness: its
setup() inserted 3 open instances sharing one rule_id, which V13 blocked
on arrival. Migrate to one rule_id per instance (pattern already used
across other storage ITs).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes the loop on three bug classes from Plan 03 triage:
context-load regressions (missing @Autowired), UI/backend drift
on template variables, and hand-maintained TS enum unions caused
by springdoc polymorphic schema quirk.
Covers 5 tasks: context-startup smoke test, template-variables
SSOT endpoint, second Playwright spec, String-to-enum migrations
on 5 condition fields, and @DiscriminatorMapping on AlertCondition.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The rule editor wizard reset the condition payload on kind-change without
seeding a fireMode default; the ExchangeMatchCondition ctor allowed null to
pass through; AlertEvaluatorJob then NPE-looped every tick on a saved rule.
- core: compact ctor now rejects null fireMode (Jackson-deser path only — all
production callers already pass a value).
- V14: repair existing EXCHANGE_MATCH rows with fireMode=null to
PER_EXCHANGE + perExchangeLingerSeconds=300 (default matches the wizard).
- ui: ConditionStep.onKindChange seeds EXCHANGE_MATCH defaults so the
Select's displayed fallback ("Per exchange") is actually in form state.
- ui: validateStep('condition', ...) now enforces fireMode presence + the
mode-specific fields before the user reaches Review.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Follow-up to 83837ada addressing the critical-review feedback:
- Duplicate ConditionKind type consolidated: the one in
api/queries/alertRules.ts (which was nullable — wrong) is gone;
single source of truth lives in this module.
- Module moved out of api/ into pages/Alerts/ where it belongs.
api/ is the data layer; labels + hide lists are view-layer concerns.
- Hidden values formalised: Comparator.EQ and JvmAggregation.LATEST
are intentionally not surfaced in dropdowns (noisy / wrong feature
boundary, see in-file comments). They remain in the type unions so
rules that carry those values save/load correctly — we just don't
advertise them in the UI.
- JvmAggregation declaration order restored to MAX/AVG/MIN (matches
what users saw before 83837ada). LATEST declared last; hidden.
- Snapshot tests for every visible *_OPTIONS array — reviewer signal
in future PRs when a backend enum change or hide-list edit
silently reshapes the dropdown.
- `toOptions` gains a JSDoc noting that label-map declaration order
is load-bearing (ES2015 Object.keys insertion-order guarantee).
- **Honest about the springdoc schema quirk**: the generated
polymorphic condition types resolve to `never` at the TypeScript
level (two conflicting `kind` discriminators — the class-name
literal and the Jackson enum — intersect to never), which silently
defeated `Record<T, string>` exhaustiveness. The previous commit's
"schema-derived enums" claim was accurate only for the flat-field
enums (ConditionKind, Severity, TargetKind); condition-specific
enums (RouteMetric, Comparator, JvmAggregation, ExchangeFireMode)
were silently `never`. Those are now declared as hand-written
string-literal unions with a top-of-file comment spelling out the
issue and the regen-and-compare workflow. Real upstream fix is a
backend-side adjustment to how springdoc emits polymorphic
`@JsonSubTypes` — out of scope for this phase.
Verified: ui build green, 56/56 vitest pass (49 pre-existing + 7
new enum snapshots).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes item 5 on the Plan 03 cleanup triage. The option arrays
("METRICS", "COMPARATORS", KIND_OPTIONS, SEVERITY_OPTIONS, FIRE_MODES)
scattered across RouteMetricForm / JvmMetricForm / ExchangeMatchForm /
ConditionStep / ScopeStep were hand-typed string literals. They drifted
silently — P95_LATENCY_MS appeared in a dropdown without a backend
counterpart (caught at runtime in bcde6678); JvmMetric.LATEST and
Comparator.EQ existed on the backend but were missing from the UI all
along.
Fix: new `ui/src/api/alerting-enums.ts` derives every enum from
schema.d.ts and pairs each with a `Record<T, string>` label map.
TypeScript enforces exhaustiveness — adding or removing a backend
value fails the build of this file until the label map is updated.
Every consumer imports the generated `*_OPTIONS` array.
Covered (schema-derived):
- ConditionKind → CONDITION_KIND_OPTIONS
- Severity → SEVERITY_OPTIONS
- RouteMetric → ROUTE_METRIC_OPTIONS
- Comparator → COMPARATOR_OPTIONS (adds EQ that was missing)
- JvmAggregation → JVM_AGGREGATION_OPTIONS (adds LATEST that was missing)
- ExchangeMatch.fireMode → EXCHANGE_FIRE_MODE_OPTIONS
- AlertRuleTarget.kind → TARGET_KIND_OPTIONS
form-state.ts: `severity: 'CRITICAL' | 'WARNING' | 'INFO'` and
`kind: 'USER' | 'GROUP' | 'ROLE'` literal unions swapped for the
derived `Severity` / `TargetKind` aliases.
Not covered, backend types them as `String` (no `@Schema(allowableValues)`
annotation yet):
- AgentStateCondition.state
- DeploymentStateCondition.states
- LogPatternCondition.level
- ExchangeFilter.status
- JvmMetricCondition.metric
These stay hand-typed with a pointer-comment. Follow-up: add
`@Schema(allowableValues = …)` to the Java record components so the
enums land in schema.d.ts; then fold them into alerting-enums.ts.
Plus: gitnexus index-stats refresh in AGENTS.md/CLAUDE.md from the
post-deploy reindex.
Verified: ui build green, 49/49 vitest pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After the UiAuth / Oidc / UserAdmin controllers were aligned to store
bare user_ids, the rules that future sessions read were still describing
the old behaviour (OutboundConnectionAdminController "strips user:
prefix" — true mechanically but the subtlety is that the strip is
the bridge between a prefixed JWT subject and an unprefixed DB key,
not a hack).
- CLAUDE.md: expand the User persistence one-liner to state the
convention authoritatively (local `<username>`, OIDC `oidc:<sub>`,
JWT `user:` namespace, env-scoped controllers strip for FK).
- .claude/rules/app-classes.md:
- Add "User ID conventions" section near the top that spells out
write-path vs read-path behaviour in one place.
- Add UiAuthController + OidcAuthController entries under
security/ with their upsert shape documented.
- Soften the OutboundConnectionAdminController line to reference
the convention instead of restating the mechanism.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Follow-up to the UiAuthController fix: every write path that puts a row
into users/user_roles/user_groups must use the bare DB key, because
the env-scoped controllers (Alert, AlertRule, AlertSilence, Outbound)
strip "user:" before using the name as an FK. If the write path stores
prefixed, first-time alerting/outbound writes fail with
alert_rules_created_by_fkey violation.
UiAuthController shipped the model in the prior commit (bare userId
for all DB/RBAC calls, "user:"-namespaced subject for JWT signing).
Bringing the other two write paths in line:
- OidcAuthController.callback:
userId = "oidc:" + oidcUser.subject() // DB key, no "user:"
subject = "user:" + userId // JWT subject (namespaced)
All userRepository / rbacService / applyClaimMappings calls use
userId. Tokens still carry the namespaced subject so
JwtAuthenticationFilter can distinguish user vs agent tokens.
- UserAdminController.createUser: userId = request.username() (bare).
resetPassword: dropped the "user:"-strip fallback that was only
needed because create used to prefix — now dead.
No migration. Greenfield alpha product — any pre-existing prefixed
rows in a dev DB will become orphans on next login (login upserts
the unprefixed row, old prefixed row is harmless but unused).
Operators doing a clean re-index can wipe the DB.
Read-path controllers still strip — harmless for bare DB rows, and
OIDC humans (JWT sub "user:oidc:<s>") still resolve correctly to
the new DB key "oidc:<s>" after stripping.
Verified: 45/45 alerting + outbound ITs pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Post-commit stats: 8524 nodes / 22174 edges / 415 clusters / 300 flows
(up from 8513/22146/409/300 after the five cleanup-batch commits).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Root cause of the mismatch that prompted the one-shot cameleer-seed
docker service: UiAuthController stored users.user_id as the JWT
subject "user:admin" (JWT sub format). Every env-scoped controller
(Alert, AlertSilence, AlertRule, OutboundConnectionAdmin) already
strips the "user:" prefix on the read path — so the rest of the
system expects the DB key to be the bare username. With UiAuth
storing prefixed, fresh docker stacks hit
"alert_rules_created_by_fkey violation" on the first rule create.
Fix: inside login(), compute `userId = request.username()` and use
it everywhere the DB/RBAC layer is touched (isLocked, getPasswordHash,
record/clearFailedLogins, upsert, assignRoleToUser, addUserToGroup,
getSystemRoleNames). Keep `subject = "user:" + userId` — we still
sign JWTs with the namespaced subject so JwtAuthenticationFilter can
distinguish user vs agent tokens.
refresh() and me() follow the same rule via a stripSubjectPrefix()
helper (JWT subject in, bare DB key out).
With the write path aligned, the docker bridge is no longer needed:
- Deleted deploy/docker/postgres-init.sql
- Deleted cameleer-seed service from docker-compose.yml
Scope: UiAuthController only. UserAdminController + OidcAuthController
still prefix on upsert — that's the bug class the triage identified
as "Option A or B either way OK". Not changing them now because:
a) prod admins are provisioned unprefixed through some other path,
so those two files aren't the docker-only failure observed;
b) stripping them would need a data migration for any existing
prod users stored prefixed, which is out of scope for a cleanup
phase. Follow-up worth scheduling if we ever wire OIDC or admin-
created users into alerting FKs.
Verified: 33/33 alerting+outbound controller ITs pass (9 outbound,
10 rules, 9 silences, 5 alert inbox).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Spec §13 calls for the notification bell to colour-code by highest
unread severity (CRITICAL → error, WARNING → amber, INFO → muted).
The old { count } DTO forced the UI to pick one static colour, so
NotificationBell shipped with a TODO. Grow the contract instead:
UnreadCountResponse = { total, bySeverity: { CRITICAL, WARNING, INFO } }
Guarantees:
- every severity is always present with a >=0 value (no undefined
keys on the wire), so the UI can branch without defaults.
- total = sum of bySeverity values — kept explicit on the wire for
cheap top-line display, not recomputed client-side.
Backend
- AlertInstanceRepository: replaces countUnreadForUser(long) with
countUnreadBySeverityForUser returning Map<AlertSeverity, Long>.
One SQL round-trip per (env, user) — GROUP BY ai.severity over the
same NOT EXISTS(alert_reads) filter.
- UnreadCountResponse.from(Map) normalises and defensively copies;
missing severities default to 0.
- InAppInboxQuery.countUnread now returns the DTO, caches the full
response (still 5s TTL) so severity breakdown gets the same
hit-rate as the total did before.
- AlertController just hands the DTO back.
Breaking change — no backwards-compat shim: the `count` field is
gone. UI and tests updated in the same commit; there are no other
API consumers in the tree.
Frontend
- Regenerated openapi.json + schema.d.ts against a fresh build of
the new backend.
- NotificationBell branches badge colour on the highest unread
severity (CRITICAL > WARNING > INFO) via new CSS variants.
- Tests cover all four paths: zero, critical-present, warning-only,
info-only.
Tests: 7 unit tests + 12 ITs (incl. new grouping + empty-map)
+ 49 vitest (was 46; +3 severity-branch assertions).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Design spec and Plan 02 described AlertCondition polymorphism as
Id.DEDUCTION, but the code that shipped in PR #140 uses Id.NAME with
property="kind" and include=EXISTING_PROPERTY. The `kind` field is
real on every subtype and the DB stores it in a separate column
(condition_kind), so reading the discriminator directly is simpler
than deduction — update the docs to match. Also add `"kind"` to the
example JSON payloads so they match on-wire reality.
OutboundAuth (Plan 01) correctly still uses Id.DEDUCTION and is
unchanged.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Added as a reusable primitive during Plan 03 Task 9, but the intended
consumer (NotificationBell live-region refresh) was removed during
code review, leaving the hook unused. Delete it — YAGNI; reintroduce
when a real consumer shows up.
Verified upstream impact (gitnexus): 0 callers, LOW risk.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
fixtures.ts: auto-applied login fixture — visits /login?local to skip OIDC
auto-redirect, fills username/password via label-matcher, clicks 'Sign in',
then selects the 'default' env so alerting hooks enable (useSelectedEnv gate).
Override via E2E_ADMIN_USER + E2E_ADMIN_PASS.
alerting.spec.ts: 4 tests against the full docker-compose stack:
- sidebar Alerts accordion → /alerts/inbox
- 5-step wizard: defaults-only create + row delete (unique timestamp name
avoids strict-mode collisions with leftover rules)
- CMD-K palette via SearchTrigger click (deterministic; Ctrl+K via keyboard
is flaky when the canvas doesn't have focus)
- silence matcher-based create + end-early
DS FormField renders labels as generics (not htmlFor-wired), so inputs are
targeted by placeholder or label-proximity locators instead of getByLabel.
Does not exercise fire→ack→clear; that's covered backend-side by
AlertingFullLifecycleIT (Plan 02). UI E2E for that path would need event
injection into ClickHouse, out of scope for this smoke.
Alerting + outbound controllers resolve acting user via
authentication.name with 'user:' prefix stripped → 'admin'. But
UserRepository.upsert stores env-admin as 'user:admin' (JWT sub format).
The resulting FK mismatch manifests as 500 'alert_rules_created_by_fkey'
on any create operation in a fresh docker stack.
Workaround: run-once 'cameleer-seed' compose service runs psql against
deploy/docker/postgres-init.sql after the server is healthy (i.e. after
Flyway migrations have created tenant_default.users), inserting
user_id='admin' idempotently. The root-cause fix belongs in the backend
(either stop stripping the prefix in alerting/outbound controllers, or
normalise storage to the unprefixed form) and is out of scope for
Plan 03.
- RouteMetricForm dropped P95_LATENCY_MS — not in cameleer-server-core
RouteMetric enum (valid: ERROR_RATE, P99_LATENCY_MS, AVG_DURATION_MS,
THROUGHPUT, ERROR_COUNT).
- initialForm now returns a ready-to-save ROUTE_METRIC condition
(metric=ERROR_RATE, comparator=GT, threshold=0.05, windowSeconds=300),
so clicking through the wizard with all defaults produces a valid rule.
Prevents a 400 'missing type id property kind' + 400 on condition enum
validation if the user leaves the condition step untouched.
Task 29's refactor added a package-private test-friendly constructor
alongside the public production one. Without @Autowired Spring cannot pick
which constructor to use for the @Component, and falls back to searching
for a no-arg default — crashing startup with 'No default constructor found'.
Detected when launching the server via the new docker-compose stack; unit
tests still pass because they invoke the package-private test constructor
directly.
Mirrors the k8s manifests in deploy/ as a local dev stack:
- cameleer-postgres (matches deploy/cameleer-postgres.yaml)
- cameleer-clickhouse (matches deploy/cameleer-clickhouse.yaml, default CLICKHOUSE_DB=cameleer)
- cameleer-server (built from Dockerfile, env mirrors deploy/base/server.yaml)
- cameleer-ui (built from ui/Dockerfile, served on host :8080 to leave :5173 free for Vite dev)
Dockerfile + ui/Dockerfile: REGISTRY_TOKEN is now optional (empty → skip Maven/npm auth).
cameleer-common package is public, so anonymous pulls succeed; private packages still require the token.
Backend defaults tuned for local E2E:
- RUNTIME_ENABLED=false (no Docker-in-Docker deployments in dev stack)
- OUTBOUND_HTTP_ALLOW_PRIVATE_TARGETS=true (so webhook tests can target host.docker.internal etc.)
- UIUSER/UIPASSWORD=admin/admin (matches Playwright E2E_ADMIN_USER/PASS defaults)
- CORS includes both :5173 (Vite) and :8080 (nginx)
.claude/rules/ui.md now maps every Plan 03 UI surface. Admin guide gains
an inbox/rules/silences walkthrough so ops teams can start in the UI
without reading the spec.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Prometheus scrapes can fire every few seconds. The open-alerts / open-rules
gauges query Postgres on each read — caching the values for 30s amortises
that to one query per half-minute. Addresses final-review NIT from Plan 02.
- Introduces a package-private TtlCache that wraps a Supplier<Long> and
memoises the last read for a configurable Duration against a Supplier<Instant>
clock.
- Wraps each gauge supplier (alerting_rules_total{enabled|disabled},
alerting_instances_total{state}) in its own TtlCache.
- Adds a test-friendly constructor (package-private) taking explicit
Duration + Supplier<Instant> so AlertingMetricsCachingTest can advance
a fake clock without waiting wall-clock time.
- Adds AlertingMetricsCachingTest covering:
* supplier invoked once per TTL across repeated scrapes
* 29 s elapsed → still cached; 31 s elapsed → re-queried
* gauge value reflects the cached result even after delegate mutates
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Rejects webhook URLs that resolve to loopback, link-local, or RFC-1918
private ranges (IPv4 + IPv6 ULA fc00::/7). Enforced on both create and
update in OutboundConnectionServiceImpl before persistence; returns 400
Bad Request with "private or loopback" in the body.
Bypass via `cameleer.server.outbound-http.allow-private-targets=true`
for dev environments where webhooks legitimately point at local
services. Production default is `false`.
Test profile sets the flag to `true` in application-test.yml so the
existing ITs that post webhooks to WireMock on https://localhost:PORT
keep working. A dedicated OutboundConnectionSsrfIT overrides the flag
back to false (via @TestPropertySource + @DirtiesContext) to exercise
the reject path end-to-end through the admin controller.
Plan 01 scope; required before SaaS exposure (spec §17).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Extends operationalSearchData with open alerts (FIRING|ACKNOWLEDGED) and
all rules. Badges convey severity + state. Selecting an alert navigates to
/alerts/inbox/{id}; a rule navigates to /alerts/rules/{id}. Uses the
existing CommandPalette extension point — no new registry.
Fetches target-env apps (useCatalog) and env-allowed outbound
connections, passes them to prefillFromPromotion, and renders the
returned warnings in an amber banner above the step nav. Warnings list
the field name and the remediation message so users see crossings that
need manual adjustment before saving.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Review step dumps a human summary plus raw request JSON, and (when a
setter is supplied) offers an Enabled-on-save Toggle. Promotion prefill
now returns {form, warnings}: clears agent IDs (per-env), flags missing
apps in target env, and flags webhook connections not allowed in target
env. 4 Vitest cases cover copy-name, agent clear, app-missing, and
webhook-not-allowed paths.
The wizard now consumes {form, warnings}; Task 25 renders the warnings
banner.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Title and message use MustacheEditor with kind-specific autocomplete.
Preview button posts to the render-preview endpoint and shows rendered
title/message inline. Targets combine users/groups/roles into a unified
Badge pill list. Webhook picker filters to outbound connections allowed
in the current env (spec 6, allowed_environment_ids). Header overrides
use plain Input rather than MustacheEditor for now.
Deviations:
- RenderPreviewRequest is Record<string, never>, so we send {} instead
of {titleTemplate, messageTemplate}; backend resolves from rule state.
- RenderPreviewResponse has {title, message} (plan draft used
renderedTitle/renderedMessage).
- Button size="sm" not "small" (DS only accepts sm|md).
- Target kind field renamed from targetKind to kind to match
AlertRuleTarget DTO.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three numeric inputs for evaluation cadence, for-duration, and
re-notification window, plus a Test evaluate button for saved rules.
TestEvaluateRequest is empty on the wire (server uses the rule id), so
we send {} and rely on the backend to evaluate the current saved state.
Deviation: plan draft passed {condition: toRequest(form).condition} into
the request body. The generated TestEvaluateRequest type is
Record<string, never>, so we send an empty body.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Each condition kind (ROUTE_METRIC, EXCHANGE_MATCH, AGENT_STATE,
DEPLOYMENT_STATE, LOG_PATTERN, JVM_METRIC) renders its own payload-shape
form. Changing the kind resets the condition payload to {kind, scope} so
stale fields from a previous kind don't leak into the save request.
Deviation: DS Select uses native event-based onChange. Plan draft showed
a value-based signature (onChange(v) => ...).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Name, description, severity, scope-kind radio, and cascading app/route/
agent selectors driven by catalog + agents data. Adjusts condition
routing by clearing routeId/agentId when the app changes.
Deviation: DS Select uses native event-based onChange; plan draft had
a value-based signature.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wizard navigates 5 steps (scope/condition/trigger/notify/review) with
per-step validation. form-state module is the single source of truth for
the rule form; initialForm/toRequest/validateStep are unit-tested (6
tests). Step components are stubbed and will be implemented in Tasks
20-24. prefillFromPromotion is a thin wrapper in this commit; Task 24
rewrites it to compute scope-adjustment warnings.
Deviation notes:
- FormState.targets uses {kind, targetId} to match AlertRuleTarget DTO
field names (plan draft had targetKind).
- toRequest casts through Record<string, unknown> so the spread over
the Partial<AlertCondition> union typechecks.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Promotion dropdown builds a /alerts/rules/new URL with promoteFrom, ruleId,
and targetEnv query params — the wizard will read these in Task 24 and
pre-fill the form with source-env prefill + client-side warnings.
AlertRow is reused by AllAlertsPage and HistoryPage. Marking a row as read
happens when its link is followed (the detail sub-route will be added in
phase 10 polish). FIRING rows get an amber left border.
Renders the `<NotificationBell />` as the first child of `<TopBar>` (before
`<SearchTrigger>`). The bell links to `/alerts/inbox` and shows the unread
alert count for the currently selected environment.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds `buildAlertsTreeNodes` to sidebar-utils and renders an Alerts section
between Applications and Starred in LayoutShell. The section uses an
accordion pattern — entering `/alerts/*` collapses apps/admin/starred and
restores their state on leave.
gitnexus_impact(LayoutContent, upstream) = LOW (0 direct callers; rendered
only by LayoutShell's provider wrapper).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds 6 lazy-loaded route entries for the alerting UI (Inbox, All, History,
Rules list, Rule editor wizard, Silences) plus an `/alerts` → `/alerts/inbox`
redirect. Page components are placeholder stubs to be replaced in Phase 5/6/7.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wires the mustache-completion source and mustache-linter into a CodeMirror 6
EditorView. Accepts kind (filters variables) and reducedContext (env-only for
connection URLs). singleLine prevents newlines for URL/header fields. Host
ref syncs when the parent replaces value (promotion prefill).
completion fires after {{ and narrows as the user types; apply() closes the
tag automatically. Linter raises an error on unclosed {{, a warning on
references that aren't in the allowed-variable set for the current condition
kind. Kind-specific allowed set comes from availableVariables().
Plan prose had spec §8 idealized leaves, but the backend NotificationContext
only emits a subset:
ROUTE_METRIC / EXCHANGE_MATCH → route.id + route.uri (uri added)
LOG_PATTERN → log.pattern + log.matchCount (renamed from log.logger/level/message)
app.slug / app.id → scoped to non-env kinds (removed from 'always')
exchange.link / alert.comparator / alert.window / app.displayName → removed (backend doesn't emit)
Without this alignment the Task 11 linter would (1) flag valid route.uri as
unknown, (2) suggest log.{logger,level,message} as valid paths that render
empty, and (3) flag app.slug on env-wide rules.
ALERT_VARIABLES mirrors the spec §8 context map. availableVariables(kind)
returns the kind-specific filter (always vars + kind vars). extractReferences
+ unknownReferences drive the inline amber linter. Backend NotificationContext
adds must land here too.
NotificationBell used a usePageVisible() subscription that re-rendered on
every visibilitychange without consuming the value. TanStack Query's
refetchIntervalInBackground:false already pauses polling; the extra
subscription was speculative generality. Dropped the import + call + JSDoc
reference; usePageVisible hook + test retained as a reusable primitive.
Also: alerts.test.tsx 'returns the server payload unmodified' asserted a
pre-plan {total, bySeverity} shape, but UnreadCountResponse is actually
{count}. Fixed mock + assertion to {count: 3}.
Adds a header bell component linking to /alerts/inbox with an unread-count
badge for the selected environment. Polling pauses when the tab is hidden
via TanStack Query's refetchIntervalInBackground:false (already set on
useUnreadCount); the new usePageVisible hook gives components a
re-renders-on-visibility-change signal for future defense-in-depth.
Plan-prose deviation: the plan assumed UnreadCountResponse carries a
bySeverity map for per-severity badge coloring, but the backend DTO only
exposes a scalar `count`. The bell reads `data?.count` and renders a single
var(--error) tint; a TODO references spec §13 for future per-severity work
that would require expanding the DTO.
Tests: usePageVisible toggles on visibilitychange events; NotificationBell
renders the bell with no badge at count=0 and shows "3" at count=3.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
State colors follow the convention from @cameleer/design-system (CRITICAL->error,
WARNING->warning, INFO->auto). Silenced pill stacks next to state for the spec
section 8 audit-trail surface.
Adds env-scoped hooks for the alerts inbox:
- useAlerts (30s poll, background-paused, filter-aware)
- useAlert, useUnreadCount (30s poll)
- useAckAlert, useMarkAlertRead, useBulkReadAlerts (mutations that
invalidate the alerts query key tree + unread-count)
Test file uses .tsx because the QueryClientProvider wrapper relies on
JSX; vitest picks up both .ts and .tsx via the configured include glob.
Client mock targets the actual export name (`api` in ../client) rather
than the `apiClient` alias that alertMeta re-exports.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fetched from http://192.168.50.86:30090/api/v1/api-docs via
`npm run generate-api:live`. Adds TypeScript types for the new alerting
REST surface merged in #140:
- 15 alerting paths under /environments/{envSlug}/alerts/** (rules CRUD,
enable/disable, render-preview, test-evaluate, inbox, unread-count,
ack/read/bulk-read, silences CRUD, per-alert notifications)
- 1 flat notification retry path /alerts/notifications/{id}/retry
- 4 outbound-connection admin paths (from Plan 01 #139)
Verified tsc -p tsconfig.app.json --noEmit exits 0 — no existing SPA
call sites break against the fresh types. Plan 03 UI work can consume
these directly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Install @codemirror/{view,state,autocomplete,commands,language,lint}
and @lezer/common — needed by Phase 3's MustacheEditor (Task 13).
CM6 picked over a raw textarea for its small incremental-rendering
bundle, full ARIA/keyboard support, and pluggable autocomplete +
linter APIs that map cleanly to Mustache token parsing.
Add ui/playwright.config.ts wiring Task 30's E2E smoke:
- testDir ./src/test/e2e, single worker, trace+screenshot on failure
- webServer launches `npm run dev:local` (backend on :8081 required)
- PLAYWRIGHT_BASE_URL env var skips the dev server for CI against a
pre-deployed UI
Add test:e2e / test:e2e:ui npm scripts and exclude Playwright's
test-results/ and playwright-report/ from git. @playwright/test
itself was already in devDependencies from an earlier task.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Prepares for Plan 03 unit tests (MustacheEditor, NotificationBell, wizard step
validation). jsdom environment + jest-dom matchers + canary test verifies the
wiring.
Fetched from http://192.168.50.86:30090/api/v1/api-docs via
`npm run generate-api:live`. Adds TypeScript types for the new alerting
REST surface merged in #140:
- 15 alerting paths under /environments/{envSlug}/alerts/** (rules CRUD,
enable/disable, render-preview, test-evaluate, inbox, unread-count,
ack/read/bulk-read, silences CRUD, per-alert notifications)
- 1 flat notification retry path /alerts/notifications/{id}/retry
- 4 outbound-connection admin paths (from Plan 01 #139)
Verified tsc -p tsconfig.app.json --noEmit exits 0 — no existing SPA
call sites break against the fresh types. Plan 03 UI work can consume
these directly.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Follow-up to #141. AbstractPostgresIT centrally declared three @MockBean
fields (clickHouseSearchIndex, clickHouseLogStore, agentRegistryService),
which meant EVERY IT ran against mocks instead of the real Spring context.
That masked the production crashloop — the real bean graph was never
exercised by CI.
- Remove the three @MockBean fields from AbstractPostgresIT.
- Move @MockBean declarations onto only the specific ITs that stub
method behavior (verified by grepping for when/verify calls).
- ITs that don't stub CH behavior now inject the real beans.
- Add SpringContextSmokeIT — @SpringBootTest with no mocks, void
contextLoads(). Fails fast on declared-type / autowire-type
mismatches like the one #141 fixed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Production crashlooped on startup: ExchangeMatchEvaluator autowires the
concrete ClickHouseSearchIndex (for countExecutionsForAlerting, which
lives only on the concrete class, not the SearchIndex interface), but
StorageBeanConfig declared the bean with interface return type SearchIndex.
Spring matches autowire candidates by declared bean type, not by runtime
instance class, so the concrete-typed autowire failed with:
Parameter 0 of constructor in ExchangeMatchEvaluator required a bean
of type 'ClickHouseSearchIndex' that could not be found.
ClickHouseLogStore's bean is already declared with the concrete return
type (line 171), which is why LogPatternEvaluator autowires fine.
All alerting ITs passed pre-merge because AbstractPostgresIT replaces the
clickHouseSearchIndex bean with @MockBean(name=...) whose declared type
IS the concrete ClickHouseSearchIndex. The mock masked the prod bug.
Follow-up: remove @MockBean(name="clickHouseSearchIndex") from
AbstractPostgresIT so the real bean graph is exercised by alerting ITs
(and add a SpringContextSmokeIT that loads the context with no mocks).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Documents the three Flyway migrations added by the alerting feature branch
so future sessions have an accurate migration map.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Rule creation now goes through POST /alerts/rules (exercises saveTargets on the
write path). Clock is replaced with @MockBean(name="alertingClock") and re-stubbed
in @BeforeEach to survive Mockito's inter-test reset. Six ordered steps:
1. seed log → tick evaluator → assert FIRING instance with non-empty targets (B-1)
2. tick dispatcher → assert DELIVERED notification + lastNotifiedAt stamped (B-2)
3. ack via REST → assert ACKNOWLEDGED state
4. create silence → inject PENDING notification → tick dispatcher → assert silenced (FAILED)
5. delete rule → assert rule_id nullified, rule_snapshot preserved (ON DELETE SET NULL)
6. new rule with reNotifyMinutes=1 → first dispatch → advance clock 61s →
evaluator sweep → second dispatch → verify 2 WireMock POSTs (B-2 cadence)
Background scheduler races addressed by resetting claimed_by/claimed_until before
each manual tick. Simulated clock set AFTER log insert to guarantee log timestamp
falls within the evaluator window. Re-notify notifications backdated in Postgres
to work around the simulated vs real clock gap in claimDueNotifications.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
newInstance() now maps rule.targets() into targetUserIds/targetGroupIds/targetRoleNames
so newly created AlertInstance rows carry the correct target arrays.
Previously these were always empty List.of(), making the inbox query return nothing.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
V13 migration creates alert_instances_open_rule_uq — a partial unique index on
(rule_id) WHERE state IN ('PENDING','FIRING','ACKNOWLEDGED'), preventing
duplicate open instances per rule. PostgresAlertInstanceRepository.save() catches
DuplicateKeyException and returns the existing open instance instead of failing.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
AlertNotificationRepository gains resetForRetry(UUID, Instant) which sets
attempts=0, status=PENDING, next_attempt_at=now, and clears claim/response
fields. AlertNotificationController calls resetForRetry instead of
scheduleRetry so a manual retry always starts from a clean slate.
AlertNotificationControllerIT adds retryResetsAttemptsToZero to verify
attempts==0 and status==PENDING after three prior markFailed calls.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
AlertInstanceRepository gains listFiringDueForReNotify(Instant) — only returns
instances where last_notified_at IS NOT NULL and cadence has elapsed (IS NULL
branch excluded: sweep only re-notifies, initial notify is the dispatcher's job).
AlertEvaluatorJob.sweepReNotify() runs at the end of each tick, enqueues fresh
notifications for eligible instances and stamps last_notified_at.
NotificationDispatchJob stamps last_notified_at on the alert_instance when a
notification is DELIVERED, providing the anchor timestamp for cadence checks.
PostgresAlertInstanceRepositoryIT adds listFiringDueForReNotify test covering
the three-rule eligibility matrix (never-notified, long-ago, recent).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
saveTargets() is called unconditionally at the end of save() — it deletes
existing targets and re-inserts from the current targets list. findById()
and listByEnvironment() already call withTargets() so reads are consistent.
PostgresAlertRuleRepositoryIT adds saveTargets_roundtrip and
saveTargets_updateReplacesExistingTargets to cover the new write path.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Investigated three approaches for CH 24.12:
- Inline SETTINGS on ADD PROJECTION: rejected (UNKNOWN_SETTING — not a query-level setting).
- ALTER TABLE MODIFY SETTING deduplicate_merge_projection_mode='rebuild': works; persists in
table metadata across connection restarts; runs before ADD PROJECTION in the SQL script.
- Session-level JDBC URL param: not pursued (MODIFY SETTING is strictly better).
alerting_projections.sql now runs MODIFY SETTING before the two executions ADD PROJECTIONs.
AlertingProjectionsIT strengthened to assert all four projections (including alerting_app_status
and alerting_route_status on executions) exist after schema init.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The evaluator mapped P95_LATENCY_MS to ExecutionStats.avgDurationMs because
stats_1m_route has no p95 column. Exposing the old name implied p95 semantics
operators did not get. Rename to AVG_DURATION_MS makes the contract honest.
Updated RouteMetric enum (with javadoc), evaluator switch, and admin guide.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
AbstractPostgresIT gained clickHouseSearchIndex and agentRegistryService mocks in Phase 9.
All 14 alerting IT subclasses that re-declared the same @MockBean fields now fail with
"Duplicate mock definition". Removed the redundant declarations; per-class clickHouseLogStore
mock kept where needed. 120 alerting tests now pass (0 failures).
Also adds docs/alerting-02-verification.md (Task 43).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds alerting stanza to application.yml with all AlertingProperties
fields backed by env-var overrides. Creates docs/alerting.md covering
six condition kinds (with example JSON), template variables, webhook
setup (Slack/PagerDuty examples), silence patterns, circuit-breaker
and retention troubleshooting, and Prometheus metrics reference.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Nightly @Scheduled(03:00) job deletes RESOLVED alert_instances older
than eventRetentionDays and DELIVERED/FAILED alert_notifications older
than notificationRetentionDays. Uses injected Clock for testability.
IT covers: old-resolved deleted, fresh-resolved kept, FIRING kept
regardless of age, PENDING notification never deleted.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- GET /environments/{envSlug}/alerts/{alertId}/notifications — list notifications for instance (VIEWER+)
- POST /alerts/notifications/{id}/retry — manual retry of failed notification (OPERATOR+)
Flat path because notification IDs are globally unique (no env routing needed)
- scheduleRetry resets attempts to 0 and sets nextAttemptAt = now
- Added 11 alerting path matchers to SecurityConfig before outbound-connections block
- Fixed context loading failure in 6 pre-existing alerting storage/migration ITs by adding
@MockBean(clickHouseSearchIndex/clickHouseLogStore): ExchangeMatchEvaluator and
LogPatternEvaluator inject the concrete classes directly (not interface beans), so the
full Spring context fails without these mocks in tests that don't use the real CH container
- 5 IT tests: list, viewer-can-list, retry, viewer-cannot-retry, unknown-404
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- POST/GET/DELETE /environments/{envSlug}/alerts/silences
- 422 when endsAt <= startsAt ("endsAt must be after startsAt")
- OPERATOR+ for create/delete, VIEWER+ for list
- Audit: ALERT_SILENCE_CREATE/DELETE with AuditCategory.ALERT_SILENCE_CHANGE
- 6 IT tests: create, viewer-list, viewer-cannot-create, bad time-range, delete, audit event
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- GET /environments/{envSlug}/alerts — inbox filtered by userId/groupIds/roleNames via InAppInboxQuery
- GET /unread-count — memoized unread count (5s TTL)
- GET /{id}, POST /{id}/ack, POST /{id}/read, POST /bulk-read
- bulkRead filters instanceIds to env before delegating to AlertReadRepository
- VIEWER+ for all endpoints; env isolation enforced by requireInstance
- 7 IT tests: list, env isolation, unread-count, ack flow, read, bulk-read, viewer access
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- POST/GET/PUT/DELETE /environments/{envSlug}/alerts/rules CRUD
- POST /{id}/enable, /{id}/disable, /{id}/render-preview, /{id}/test-evaluate
- Attribute-key validation: rejects keys not matching ^[a-zA-Z0-9._-]+$ at rule-save time
(CRITICAL: ExchangeMatchCondition attribute keys are inlined into ClickHouse SQL)
- Webhook validation: verifies outboundConnectionId exists and is allowed in env
- Null-safe notification template defaults to "" for NOT NULL DB constraint
- Fixed misleading comment in ClickHouseSearchIndex to document validation contract
- OPERATOR+ for mutations, VIEWER+ for reads
- Audit: ALERT_RULE_CREATE/UPDATE/DELETE/ENABLE/DISABLE with AuditCategory.ALERT_RULE_CHANGE
- 11 IT tests covering RBAC, SQL-injection prevention, enable/disable, audit, render-preview
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
listInbox resolves user groups+roles via RbacService.getEffectiveGroupsForUser
/ getEffectiveRolesForUser then delegates to AlertInstanceRepository.
countUnread memoized per (envId, userId) with 5s TTL via ConcurrentHashMap
using a controllable Clock. 6 unit tests covering delegation, cache hit,
TTL expiry, and isolation between users/envs.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Renders URL/headers/body with Mustache, optionally HMAC-signs the body
(X-Cameleer-Signature), supports POST/PUT/PATCH, classifies 2xx/4xx/5xx
into DELIVERED/FAILED/retry. 8 WireMock-backed IT tests including HTTPS
TRUST_ALL against WireMock self-signed cert.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
HmacSHA256 signer returning sha256=<lowercase-hex>. 5 unit tests covering
known vector, prefix, hex casing, and different secrets/bodies.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add withRuleSnapshot(Map) wither to AlertInstance (same pattern as other withers)
- Call snapshotRule(rule) + withRuleSnapshot in both applyResult (single-firing) and
applyBatchFiring paths so every persisted instance carries a non-empty JSONB snapshot
- Strip null values from the Jackson-serialized map before wrapping in the immutable
snapshot so Map.copyOf in the compact ctor does not throw NPE on nullable rule fields
- Add ruleSnapshotIsPersistedOnInstanceCreation IT: asserts name/severity/conditionKind
appear in the rule_snapshot column after a tick fires an instance
- Add historySurvivesRuleDelete IT: fires an instance, deletes the rule, asserts
rule_id IS NULL and rule_snapshot still contains the rule name (spec §5 guarantee)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- AlertEvaluatorJob implements SchedulingConfigurer; fixed-delay tick from
AlertingProperties.effectiveEvaluatorTickIntervalMs (5 s floor)
- Claim-polling via AlertRuleRepository.claimDueRules (FOR UPDATE SKIP LOCKED)
- Per-kind circuit breaker guards each evaluator; failures recorded, open kinds
skipped and rescheduled without evaluation
- Single-Firing path delegates to AlertStateTransitions; new FIRING instances
enqueue AlertNotification rows per rule.webhooks()
- Batch (PER_EXCHANGE) path creates one FIRING AlertInstance per Firing entry
- PENDING→FIRING promotion handled in applyResult via state machine
- Title/message rendered via MustacheRenderer + NotificationContextBuilder;
environment resolved from EnvironmentRepository.findById per tick
- AlertEvaluatorJobIT (4 tests): uses named @MockBean replacements for
ClickHouseSearchIndex + ClickHouseLogStore; @MockBean AgentRegistryService
drives Clear/Firing/resolve cycle without timing sensitivity
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PER_EXCHANGE returns EvalResult.Batch(List<Firing>); last Firing carries
_nextCursor (Instant) in its context map for the job to persist as
evalState.lastExchangeTs.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
SilenceMatcherService.matches() evaluates AND semantics across ruleId,
severity, appSlug, routeId, agentId constraints. Null fields are wildcards.
Scope-based constraints (appSlug/routeId/agentId) return false when rule is
null (deleted rule — scope cannot be verified). 17 unit tests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sentinel-substitution approach: unresolved {{x.y.z}} tokens are replaced
with a unique NUL-delimited sentinel before Mustache compilation, rendered
as opaque text, then post-replaced with the original {{x.y.z}} literal.
Malformed templates (unclosed {{) are caught and return the raw template.
Never throws. 9 unit tests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Declared in cameleer-server-core pom (canonical location for unit-testable
rendering without Spring) and mirrored in cameleer-server-app pom so the
app module compiles standalone without a full reactor install.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds alerting_projections.sql with four projections (alerting_app_status,
alerting_route_status on executions; alerting_app_level on logs;
alerting_instance_metric on agent_metrics). ClickHouseSchemaInitializer now
runs both init.sql and alerting_projections.sql, with ADD PROJECTION and
MATERIALIZE treated as non-fatal — executions (ReplacingMergeTree) requires
deduplicate_merge_projection_mode=rebuild which is unavailable via JDBC pool.
MergeTree projections (logs, agent_metrics) always succeed and are asserted in IT.
Column names confirmed from init.sql: logs uses 'application' (not application_id),
agent_metrics uses 'collected_at' (not timestamp). All column names match the plan.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds AlertMatchSpec record (core) and ClickHouseSearchIndex.countExecutionsForAlerting —
no FINAL, no text subqueries. Filters by tenant, env, app, route, status, time window,
and optional after-cursor. Attributes (JSON string column) use inlined JSONExtractString
key literals since ClickHouse JDBC does not bind ? placeholders inside JSON functions.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds countLogs(LogSearchRequest) — no FINAL, no cursor/sort/limit —
reusing the same WHERE-clause logic as search() for tenant, env, app,
level, q, logger, source, exchangeId, and time-range filters.
Also extends ClickHouseTestHelper with executeInitSqlWithProjections()
and makes the script runner non-fatal for ADD/MATERIALIZE PROJECTION.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements AlertInstanceRepository: save (upsert), findById, findOpenForRule,
listForInbox (3-way OR: user/group/role via && array-overlap + ANY), countUnreadForUser
(LEFT JOIN alert_reads), ack, resolve, markSilenced, deleteResolvedBefore.
Integration test covers all 9 scenarios including inbox fan-out across all
three target types. Also adds @JsonIgnoreProperties(ignoreUnknown=true) to
SilenceMatcher to suppress Jackson serializing isWildcard() as a round-trip field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces the Plan 01 stub that returned [] with a real call through
AlertRuleRepository.findRuleIdsByOutboundConnectionId. Adds AlertingBeanConfig
exposing the AlertRuleRepository bean; widens OutboundBeanConfig constructor
to inject it. Delete and narrow-envs guards now correctly block when rules
reference a connection.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements AlertRuleRepository with JSONB condition/webhooks/eval_state
serialization via ObjectMapper, UPSERT on conflict, JSONB containment
query for findRuleIdsByOutboundConnectionId, and FOR UPDATE SKIP LOCKED
claim-polling for horizontal scale.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Replace hard-coded 'u1' user_id with per-test UUID to prevent PK collision on re-runs
- Add @AfterEach null-safe cleanup for environments and users rows
- Use containsExactlyInAnyOrder for enum assertions to catch misspelled names
- Slug suffix on environment insert avoids slug uniqueness conflicts on re-runs
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- OutboundConnectionRequest compact ctor: avoid NPE if tlsTrustMode is null
(defense-in-depth alongside @NotNull Bean Validation).
- Add operatorCannotTest IT case to lock the ADMIN-only contract on
POST /{id}/test — was previously untested.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Shared Spring test context meant seeded test-admin/test-operator/test-viewer/test-alice
users persisted across IT classes, breaking FlywayMigrationIT's "users is empty" assertion.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds OutboundConnectionsPage (list view with delete), lazy route at
/admin/outbound-connections, and Outbound Connections nav node in the
admin sidebar tree. No test file created — UI codebase has no existing
test infrastructure to build on.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Types are hand-authored (matching codebase admin-query convention);
schema.d.ts regeneration deferred until backend dev server is available.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
POST /{id}/test issues a synthetic probe against the connection URL.
TLS protocol/cipher/peer-cert details stubbed for now (Plan 02 follow-up).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
New audit categories: OUTBOUND_CONNECTION_CHANGE, OUTBOUND_HTTP_TRUST_CHANGE.
Controller-level @PreAuthorize defaults to ADMIN; GETs relaxed to ADMIN|OPERATOR.
SecurityConfig permits OPERATOR GETs on /api/v1/admin/outbound-connections/**.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
V11 migration referenced users(id) as uuid, but V1 users table has
user_id as TEXT primary key. Amending V11 and the OutboundConnection
record before Task 7's integration tests catch this at Flyway startup.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds ApacheOutboundHttpClientFactory (Apache HttpClient 5) that memoizes
CloseableHttpClient instances keyed on effective TLS + timeout config, and
OutboundHttpConfig (@ConfigurationProperties) that validates trusted CA paths
at startup and exposes OutboundHttpClientFactory as a Spring bean.
TRUST_ALL mode disables both cert validation (TrustAllManager in SslContextBuilder)
and hostname verification (NoopHostnameVerifier on SSLConnectionSocketFactoryBuilder).
WireMock HTTPS integration test covers trust-all bypass, system-default PKIX rejection,
and client memoization.
OIDC audit: OidcProviderHelper and OidcTokenExchanger use Nimbus SDK's own HTTP layer
(DefaultResourceRetriever for JWKS, HTTPRequest.send() for token exchange) plus the
bespoke InsecureTlsHelper for TLS skip-verify; neither uses OutboundHttpClientFactory.
Retrofit deferred to a separate follow-up per plan §20.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Defense-in-depth per code review. DTO layer already validates HTTPS at save
time; this DB-level check guards against future code paths that might bypass
the DTO validator. Mustache template variables in the URL (e.g., {{env.slug}})
remain valid since only the scheme prefix is constrained.
First of three sequenced plans for the alerting feature. Covers:
- Cross-cutting http/ module (OutboundHttpClientFactory, SslContextBuilder,
TLS trust composition, startup validation)
- Admin-managed OutboundConnection with PG persistence, AES-GCM-encrypted
HMAC secret (resolves spec §20 item 2)
- Admin CRUD REST + test endpoint + RBAC + audit
- Admin UI page with TLS config, allowed-envs multi-select, test action
- OIDC retrofit deliberately deferred (documented in Task 4 audit)
Plan 02 (alerting backend) and Plan 03 (alerting UI) written after Plan 01
executes — lets reality inform their details, especially the secret-cipher
interface and the rules-referencing integration point.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
BL-002 / gitea#138 tracks deferred native provider types (Slack Block Kit,
PagerDuty Events v2, Teams connector) with shipped templates as a post-v1
fast-follow once usage data informs which providers matter.
Spec §13 folds in context-aware variable auto-complete for the shared
<MustacheEditor /> component used in rule editor, webhook overrides, and
outbound-connection admin. Available variables filter by condition kind.
Completion engine choice added to §20 as a planning-phase decision.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Comprehensive design spec for a confined, env-scoped alerting feature:
6 signal sources, shared env-scoped rules with RBAC-targeted notifications,
in-app inbox + webhook delivery via admin-managed outbound connections,
claim-based polling for horizontal scalability, 4 CH projections for hot-path
reads. Backlog entry BL-001 / gitea#137 tracks deferred managed-CA investigation
(reuse SaaS-layer CA handling first before building in-server storage).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
JSX attribute strings don't process JS escape sequences — "Search logs\u2026"
rendered the literal "\u2026" in the placeholder. Replaced with the actual
ellipsis character.
Also aligned .logSearchInput (Application Log search) with EventFeed's
internal search input: --bg-surface background, --border border,
mono font family, 28px height. Previously used --bg-body + --border-subtle
+ body font, which looked visibly different next to the Timeline panel.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- LogQueryController: note response shape, sort param, and that the
cursor tiebreak is the insert_id UUID column (not exchange/instance)
- AgentEventsController: cursor now carries insert_id UUID (was instanceId);
order is (timestamp DESC, insert_id DESC)
- core-classes: add AgentEventPage record; note that the non-paginated
AgentEventRepository.query(...) path has been removed
- core-classes: note LogSearchRequest.sources/levels are now List<String>
with multi-value OR semantics
Keeps the rule files in sync with the cursor-pagination + multi-select
filter work on main.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Both panels now use the same card wrapper (logStyles.logCard), header
container (logStyles.logHeader, 12px 16px padding), and DS SectionHeader
for the title. Previously Timeline rendered a custom 13px span while
Application Log used SectionHeader's uppercase style, so the two panels
side-by-side looked inconsistent.
Removes the now-orphaned .eventCard/.eventCardHeader/.sectionTitle and
.timelineCard/.timelineHeader CSS rules.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
EventFeed has its own search + filter toolbar inside the component.
Wrapping it in InfiniteScrollArea made the toolbar scroll out of
sight. Drop InfiniteScrollArea for the Timeline, give EventFeed a
bounded-height flex container (it scrolls its own .list internally),
and add an explicit 'Load older events' button for cursor
pagination. Polling always on for events (low volume).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Page 1 refetches were using the captured timeRange.end, so rows
arriving after the initial render were outside the query window and
never surfaced. When timeRange.preset is set (e.g. 'last 1h'), each
fetch now advances 'to' to Date.now() so the poll picks up new rows.
Absolute ranges are unchanged.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
EventFeed's overflow-y:auto lives on its inner .list, not the root
where className lands. Extending .flatScroll to .flatScroll * covers
nested scroll containers, and relaxing the root's height:100% (which
EventFeed sets) lets content size naturally so the outer
InfiniteScrollArea owns the single scrollbar.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LogViewer and EventFeed each apply overflow-y:auto to their root
container, which produced a nested scrollbar inside the
InfiniteScrollArea that also scrolls. A flatScroll override class
flattens the DS component so the outer InfiniteScrollArea owns the
single scrollbar — matching the IntersectionObserver sentinels that
drive infinite-load.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Same-millisecond rows were silently skipped between pages because the
log cursor had no tiebreak and the events cursor tied by instance_id
(which also collides when one instance emits multiple events within a
millisecond). Add an insert_id UUID (DEFAULT generateUUIDv4()) column
to both logs and agent_events, order by (timestamp, insert_id)
consistently, and encode the cursor as 'timestamp|insert_id'. Existing
data is materialized via ALTER TABLE MATERIALIZE COLUMN (one-time
background mutation).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reversing logStream.items client-side breaks across infinite-scroll
pages. Passing sort='asc'/'desc' into the query key and URL triggers
a fresh first-page fetch in the selected order.
AgentEventService.queryEvents, AgentEventRepository.query, and the
ClickHouse implementation have had no callers since /agents/events
became cursor-paginated. Remove them along with their dedicated IT
tests. queryPage and its tests remain as the single query path.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extends the existing ApiExceptionHandler @RestControllerAdvice to map
DateTimeParseException and IllegalArgumentException to 400 Bad Request.
Logs and agent-events endpoints both parse ISO-8601 query params and
previously leaked parse failures as internal server errors. All
IllegalArgumentException throw sites in production code are
input-validation usages (slug validation, containerConfig validation,
cursor decoding), so mapping to 400 is correct across the board.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
JDBC Timestamp binding shifted timestamps by the JVM local timezone
offset on both insert and query, producing asymmetric UTC offsets that
broke time-range filtering and cursor pagination. Switching inserts
(indexBatch, insertBufferedBatch) and all WHERE predicates to ISO-8601
strings via parseDateTime64BestEffort, and reading timestamps back as
epoch-millis via toUnixTimestamp64Milli, pins everything to UTC and
fixes the time-range filter test plus cursor pagination.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reflects LogQueryController's multi-value source/level filters,
AgentEventsController's cursor pagination shape, and the new
useInfiniteStream/InfiniteScrollArea UI primitives used by streaming
views.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Same pattern as AgentHealth, scoped to a single agent instance
(passes agentId to both log and timeline streams).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- useInfiniteStream: wrap fetchNextPage and refresh in useCallback so
InfiniteScrollArea's IntersectionObserver does not re-subscribe on
every parent render.
- InfiniteScrollArea: do not render 'End of stream' until at least one
item has loaded and the initial query has settled (was flashing on
mount before first fetch).
- AgentHealth: pass isLoading + hasItems to both InfiniteScrollArea
wrappers.
Application Log: source + level filters move server-side; text search
stays client-side. Timeline: cursor-paginated via useInfiniteAgentEvents.
Both wrapped in InfiniteScrollArea with top-gated auto-refetch.
Server-side filters on source/level/time-range, client-side text
search on top of flattened items. Leaves useApplicationLogs and
useStartupLogs untouched for bounded consumers (LogTab, StartupLogPanel).
Scrollable container with top/bottom IntersectionObserver sentinels.
Fires onTopVisibilityChange when the top is fully in view and
onEndReached when the bottom is within 100px. Used by infinite log
and event streams.
Captures the cursor-paginated /agents/events response shape
(AgentEventPageResponse with data/nextCursor/hasMore and a new ?cursor
param). Also folds in pre-existing drift from 62dd71b (environment
field on agent event rows). Consumer UI hooks are updated in
Tasks 9-11.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Wraps DateTimeParseException from Instant.parse in IllegalArgumentException
so the controller maps it to 400. Also rejects cursors with empty
instance_id (trailing '|') which would otherwise produce a vacuous
keyset predicate.
Orders by (timestamp DESC, instance_id ASC). Cursor is
base64url('timestampIso|instanceId') with a tuple keyset predicate
for stable paging across ties.
Introduces cursor-paginated query on AgentEventRepository. The cursor
format is owned by the implementation. The existing non-paginated
query(...) is kept for internal consumers.
Replaces LogSearchRequest.source (String) with sources (List<String>)
and emits 'source IN (...)' when non-empty. LogQueryController parses
?source=a,b,c the same way it parses ?level=a,b,c.
The agent_events table has an `environment` column and AgentEventsController
filters on it, but the INSERT never populated it — every row got the
column default ('default'). Result: Timeline on the Application Runtime
page was empty whenever the user's selected env was anything other than
'default'.
Thread env through the write path:
- AgentEventRepository.insert + AgentEventService.recordEvent gain an
`environment` param; delete the no-env query overload (unused).
- ClickHouseAgentEventRepository.insert writes the column (falls back to
'default' on null to match column DEFAULT).
- All 5 callers source env from the agent registry (AgentInfo.environmentId)
or the registration request body; AgentLifecycleMonitor, deregister,
command ack, event ingestion, register/re-register.
- Integration test updated for the new signatures.
Pre-existing rows in deployed CH will still report environment='default'.
New events from this build forward will carry the correct env. Backfill
(UPDATE ... FROM apps) is left as a manual DB step if historical timeline
is needed for non-default envs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LayoutShell's <main> sets overflow:hidden + min-height:0; pages must
handle their own scroll. SwaggerPage's root div had no height constraint,
so the Swagger UI rendered below the viewport with no scrollbar.
Match the pattern used by AppsTab/DashboardTab/etc.: flex:1 + min-height:0
+ overflow-y:auto on the page root. Inline style since this is a leaf
page with no existing module.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The deployed instance at 192.168.50.86:30090 is the canonical source of
truth for the API schema during development — regen against it instead
of requiring a local backend boot with Postgres + ClickHouse.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Fetched from http://192.168.50.86:30090/api/v1/api-docs (running origin/main
through b7a107d — full P3B/P3C env-scoping migration live there).
SPA TS types now match the env-scoped URL shape used at runtime:
- /environments/{envSlug}/... for data, config, search, logs, routes, agents
- /agents/config (agent-authoritative)
- /admin/environments/{envSlug}/... (env CRUD)
Note: ExecutionDetail.environment isn't in the regenerated schema yet —
commit d02fa73 (local, not yet pushed/deployed) adds that backend field.
The local type extension in ui/src/components/ExecutionDiagram/types.ts
covers the gap until the next redeploy + regen.
UI typecheck (tsc -p tsconfig.app.json --noEmit) passes.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Correlated exchanges always share the env of the one being viewed —
using the globally-selected env from the picker was wrong if the user
switched envs after opening a detail view (or arrived via permalink).
Thread `environment` through:
- `ExecutionStore.ExecutionRecord` gains `environment` field; the
ClickHouse `executions` table already stores this, just not read back.
- `ClickHouseExecutionStore.findById` SELECT adds the column; mapper
populates it.
- `ExecutionDetail` gains `environment`; `DetailService` passes through.
- `IngestionService.toExecutionRecord` passes null — this legacy PG
ingestion path isn't active when ClickHouse is enabled, and the
read-side is what drives the correlation UI.
- UI `ExchangeHeader` reads `detail.environment ?? storeEnv` and
extends the TS type locally (schema.d.ts catches up on next regen).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The env-scoping migration (P3A) changed useCorrelationChain to require an
environment arg and gate on `enabled: !!correlationId && !!environment`,
but ExchangeHeader was still calling it with one arg. Result: the query
never fired, so the header always rendered "no correlated exchanges
found" even when 4+ exchanges shared a correlationId.
Fix: read the selected env from the Zustand environment store and pass
it through.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Picks up the URL moves from P2/P3A/P3B/P3C. Also fixes a latent bug in
AppControllerIT.uploadJar_asOperator_returns201 / DeploymentControllerIT
setUp: the tests were passing the app's UUID as the {appSlug} path
variable (via `path("id").asText()`); the old AppController looked up
apps via getBySlug(), so the legacy URL call would 404 when the slug
literal was a UUID. Now the test tracks the known slug string and uses
it for every /apps/{appSlug}/... path.
Test URL updates:
- SearchControllerIT: /api/v1/search/executions →
/api/v1/environments/default/executions (GET) and
/api/v1/environments/default/executions/search (POST).
- AppControllerIT: /api/v1/apps → /api/v1/environments/default/apps.
Request bodies drop environmentId (it's in the path).
- DeploymentControllerIT: /api/v1/apps/{appId}/deployments →
/api/v1/environments/default/apps/{appSlug}/deployments. DeployRequest
body drops environmentId.
- JwtRefreshIT + RegistrationSecurityIT: smoke-test protected endpoint
call updated to the new /environments/default/executions shape.
All tests compile clean. Runtime behavior requires a full stack
(Postgres + ClickHouse + Docker); validating integration tests is a
pre-merge step before merging the feature branch.
Remaining pre-merge items (not blocked by code):
1. Regenerate ui/src/api/schema.d.ts + openapi.json by running
`cd ui && npm run generate-api:live` against a running backend.
SearchController, DeploymentController, etc. DTO signatures have
changed; schema.d.ts is frozen at the pre-migration shape.
Raw-fetch call sites introduced in P3A/P3C work at runtime without
the schema; the regen only sharpens TypeScript coverage.
2. Smoke test locally: boot server, verify EnvironmentsPage,
AppsTab, Exchanges, Dashboard, Runtime pages all function.
3. Run `mvn verify` end-to-end (Testcontainers + Docker required).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Locks the new conventions into rule files so future agents and humans
don't drift back into old patterns.
- .claude/rules/app-classes.md: replaces the flat endpoint catalog
with a taxonomy-aware reorganization (env-scoped / env-admin /
agent-only / ingestion / cross-env discovery / admin / other).
Adds the flat-endpoint allow-list with rationale per prefix and
documents the tenant-filter invariant for ClickHouse queries.
- CLAUDE.md: adds four convention bullets in Key Conventions —
URL taxonomy with allow-list pointer, slug immutability rule,
app uniqueness as (env, app_slug), env-required on env-scoped
endpoints.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P3C — the last data/query wave of the taxonomy migration. Every user-
facing read endpoint that was keyed on env-as-query-param is now under
the env-scoped URL, making env impossible to omit and unambiguous in
server-side tenant+env filtering.
Server:
- SearchController: /api/v1/search/** → /api/v1/environments/{envSlug}/...
Endpoints: /executions (GET), /executions/search (POST), /stats,
/stats/timeseries, /stats/timeseries/by-app, /stats/timeseries/by-route,
/stats/punchcard, /attributes/keys, /errors/top. Env comes from path.
- LogQueryController: /api/v1/logs → /api/v1/environments/{envSlug}/logs.
- RouteCatalogController: /api/v1/routes/catalog → /api/v1/environments/
{envSlug}/routes. Env filter unconditional (path).
- RouteMetricsController: /api/v1/routes/metrics →
/api/v1/environments/{envSlug}/routes/metrics (and /metrics/processors).
- DiagramRenderController: /{contentHash}/render stays flat (hashes are
globally unique). Find-by-route moved to /api/v1/environments/{envSlug}/
apps/{appSlug}/routes/{routeId}/diagram — the old GET /diagrams?...
handler is removed.
- Agent views split cleanly:
- AgentListController (new): /api/v1/environments/{envSlug}/agents
- AgentEventsController: /api/v1/environments/{envSlug}/agents/events
- AgentMetricsController: /api/v1/environments/{envSlug}/agents/
{agentId}/metrics — now also rejects cross-env agents (404) as a
defense-in-depth check, fulfilling B3.
Agent self-service endpoints (register/refresh/heartbeat/deregister)
remain flat at /api/v1/agents/** — JWT-authoritative.
SPA:
- queries/agents.ts, agent-metrics.ts, logs.ts, catalog.ts (route
metrics only; /catalog stays flat), processor-metrics.ts,
executions.ts (attributes/keys, stats, timeseries, search),
dashboard.ts (all stats/errors/punchcard), correlation.ts,
diagrams.ts (by-route) — all rewritten to env-scoped URLs.
- Hooks now either read env from useEnvironmentStore internally or
require it as an argument. Query keys include env so switching env
invalidates caches.
- useAgents/useAgentEvents signature simplified — env is no longer a
parameter; it's read from the store. Callers (LayoutShell,
AgentHealth, AgentInstance) updated accordingly.
- LogTab and useStartupLogs thread env through to useLogs.
- envFetch helper introduced in executions.ts for env-prefixed raw
fetch until schema.d.ts is regenerated against the new backend.
BREAKING CHANGE: All these flat paths are removed:
/api/v1/search/**, /api/v1/logs, /api/v1/routes/catalog,
/api/v1/routes/metrics (and /processors), /api/v1/diagrams
(lookup), /api/v1/agents (list), /api/v1/agents/events-log,
/api/v1/agents/{id}/metrics, /api/v1/agent-events.
Clients must use the /api/v1/environments/{envSlug}/... equivalents.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P3B of the taxonomy migration. App and deployment routes are now
env-scoped in the URL itself, making the (env, app_slug) uniqueness
key explicit. Previously /api/v1/apps/{appSlug} was ambiguous: with
the same app deployed to multiple environments (dev/staging/prod),
the handler called AppService.getBySlug(slug) which returns the
first row matching slug regardless of env.
Server:
- AppController: @RequestMapping("/api/v1/environments/{envSlug}/
apps"). Every handler now calls
appService.getByEnvironmentAndSlug(env.id(), appSlug) — 404 if the
app doesn't exist in *this* env. CreateAppRequest body drops
environmentId (it's in the path).
- DeploymentController: @RequestMapping("/api/v1/environments/
{envSlug}/apps/{appSlug}/deployments"). DeployRequest body drops
environmentId. PromoteRequest body switches from
targetEnvironmentId (UUID) to targetEnvironment (slug);
promote handler resolves the target env by slug and looks up the
app with the same slug in the target env (fails with 404 if the
target app doesn't exist yet — apps must exist in both source
and target before promote).
- AppService: added getByEnvironmentAndSlug helper; createApp now
validates slug against ^[a-z0-9][a-z0-9-]{0,63}$ (400 on
invalid).
SPA:
- queries/admin/apps.ts: rewritten. Hooks take envSlug where
env-scoped. Removed useAllApps (no flat endpoint). Renamed path
param naming: appId → appSlug throughout. Added
usePromoteDeployment. Query keys include envSlug so cache is
env-scoped.
- AppsTab.tsx: call sites updated. When no environment is selected,
the managed-app list is empty — cross-env discovery lives in the
Runtime tab (catalog). handleDeploy/handleStop/etc. pass envSlug
to the new hook signatures.
BREAKING CHANGE: /api/v1/apps/** paths removed. Clients must use
/api/v1/environments/{envSlug}/apps/{appSlug}/**. Request bodies
for POST /apps and POST /apps/{slug}/deployments no longer accept
environmentId (use the URL path instead). Promote body uses slug
not UUID.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
P3A of the taxonomy migration. Env-scoped config and settings endpoints
now live under the env-prefixed URL shape, making env a first-class
path segment instead of a query param. Agent-authoritative config is
split off into a dedicated endpoint so agent env comes from the JWT
only — never spoofable via URL.
Server:
- ApplicationConfigController: @RequestMapping("/api/v1/environments/
{envSlug}"). Handlers use @EnvPath Environment env, appSlug as
@PathVariable. Removed the dual-mode resolveEnvironmentForRead —
user flow only; agent flow moved to AgentConfigController.
- AgentConfigController (new): GET /api/v1/agents/config. Reads
instanceId from JWT subject, resolves (app, env) from registry,
returns AppConfigResponse. Registry miss → falls back to JWT env
claim for environment, but 404s if application cannot be derived
(no other source without registry).
- AppSettingsController: @RequestMapping("/api/v1/environments/
{envSlug}"). List at /app-settings, per-app at /apps/{appSlug}/
settings. Access class-wide PreAuthorize preserved (ADMIN/OPERATOR).
SPA:
- commands.ts: useAllApplicationConfigs, useApplicationConfig,
useUpdateApplicationConfig, useProcessorRouteMapping,
useTestExpression — rewritten URLs to /environments/{env}/apps/
{app}/... shape. environment now required on every call. Query
keys include environment so cache is env-scoped.
- dashboard.ts: useAppSettings, useAllAppSettings, useUpdateAppSettings
rewritten.
- TapConfigModal: new required environment prop; callers updated.
- RouteDetail, ExchangesPage: thread selectedEnv into test-expression
and modal.
Config changes in SecurityConfig for the new shape landed earlier in
P0.2; no security rule changes needed in this commit.
BREAKING CHANGE: /api/v1/config/** and /api/v1/admin/app-settings/**
paths removed. Agents must use /api/v1/agents/config instead of
GET /api/v1/config/{app}; users must use /api/v1/environments/{env}/
apps/{app}/config and /api/v1/environments/{env}/apps/{app}/settings.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
UUID-based admin paths were the only remaining UUID-in-URL pattern in
the API. Migrates /api/v1/admin/environments/{id} to /{envSlug} so
slugs are the single environment identifier in every URL. UUIDs stay
internal to the database.
- Controller: @PathVariable UUID id → @PathVariable String envSlug on
get/update/delete and the two nested endpoints (default-container-
config, jar-retention). Handlers resolve slug → Environment via
EnvironmentService.getBySlug, then delegate to existing UUID-based
service methods.
- Service: create() now validates slug against ^[a-z0-9][a-z0-9-]{0,63}$
and returns 400 on invalid slugs. Rationale documented in the class:
slugs are immutable after creation because they appear in URLs,
Docker network names, container names, and ClickHouse partition keys.
- UpdateEnvironmentRequest has no slug field and Jackson's default
ignore-unknown behavior drops any slug supplied in a PUT body;
regression test (updateEnvironment_withSlugInBody_ignoresSlug)
documents this invariant.
- SPA: mutation args change from { id } to { slug }. EnvironmentsPage
still uses env.id for local selection state (UUID from DB) but
passes env.slug to every mutation.
BREAKING CHANGE: /api/v1/admin/environments/{id:UUID}/... paths removed.
Clients must use /{envSlug}/... (slug from the environments list).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes two cross-env data leakage paths. Both endpoints previously
returned data aggregated across all environments, so a diagram or
attribute key from dev could appear in a prod UI query (and vice versa).
B1: GET /api/v1/diagrams?application=&routeId= now requires
?environment= and resolves agents via
registryService.findByApplicationAndEnvironment instead of
findByApplication. Prevents serving a dev diagram for a prod route.
B2: GET /api/v1/search/attributes/keys now requires ?environment=.
SearchIndex.distinctAttributeKeys gains an environment parameter and
the ClickHouse query adds the env filter alongside the existing
tenant_id filter. Prevents prod attribute names leaking into dev
autocompletion (and vice versa).
SPA hooks updated to thread environment through from
useEnvironmentStore; query keys include environment so React Query
re-fetches on env switch. No call-site changes needed — hook
signatures unchanged.
B3 (AgentMetricsController env scope) deferred to P3C: agent-env is
effectively 1:1 today via the instance_id naming
({envSlug}-{appSlug}-{replicaIndex}), and the URL migration in P3C
to /api/v1/environments/{env}/agents/{agentId}/metrics naturally
introduces env from path. A minimal P1 fix would regress the "view
metrics of a killed agent" case.
BREAKING CHANGE: Both endpoints now require ?environment= (slug).
Clients omitting the parameter receive 400.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Groundwork for the REST API taxonomy migration. Introduces the
infrastructure that future waves use to move data/query endpoints under
/api/v1/environments/{envSlug}/... without per-handler boilerplate.
- Add @EnvPath annotation + EnvironmentPathResolver: injects the
Environment identified by the {envSlug} path variable, 404 on unknown
slug, registered via WebConfig.addArgumentResolvers.
- Add env-scoped URL matchers to SecurityConfig (config, settings,
executions/stats/logs/routes/agents/apps/deployments under
/environments/*/**). Legacy flat matchers kept in place and will be
removed per-wave as controllers migrate. New agent-authoritative
/api/v1/agents/config matcher prepared for the agent/user split.
- Document OpenAPI schema regen workflow in CLAUDE.md so future API
changes cover schema.d.ts regeneration as part of the change.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
BREAKING: wipe dev PostgreSQL before deploying — V1 checksum changes.
Agents must now send environmentId on registration (400 if missing).
Two tables previously keyed on app name alone caused cross-environment
data bleed: writing config for (app=X, env=dev) would overwrite the row
used by (app=X, env=prod) agents, and agent startup fetches ignored env
entirely.
- V1 schema: application_config and app_settings are now PK (app, env).
- Repositories: env-keyed finders/saves; env is the authoritative column,
stamped on the stored JSON so the row agrees with itself.
- ApplicationConfigController.getConfig is dual-mode — AGENT role uses
JWT env claim (agents cannot spoof env); non-agent callers provide env
via ?environment= query param.
- AppSettingsController endpoints now require ?environment=.
- SensitiveKeysAdminController fan-out iterates (app, env) slices so each
env gets its own merged keys.
- DiagramController ingestion stamps env on TaggedDiagram; ClickHouse
route_diagrams INSERT + findProcessorRouteMapping are env-scoped.
- AgentRegistrationController: environmentId is required on register;
removed all "default" fallbacks from register/refresh/heartbeat auto-heal.
- UI hooks (useApplicationConfig, useProcessorRouteMapping, useAppSettings,
useAllAppSettings, useUpdateAppSettings) take env, wired to
useEnvironmentStore at all call sites.
- New ConfigEnvIsolationIT covers env-isolation for both repositories.
Plan in docs/superpowers/plans/2026-04-16-environment-scoping.md.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Picks up the Sidebar fix that aligns Sidebar.Section header and
Sidebar.FooterLink icon/label columns. Bottom-positioned admin section
and footer links (e.g. API Docs) now share the same 9px left offset and
16px icon wrapper width.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The exchange search silently filtered by the in-memory agent registry's
current instance IDs on top of application_id. Historical exchanges written
by previous agent instances (or any instance not currently registered, e.g.
after a server restart before agents heartbeat back) were hidden from
results even though they matched the application filter.
Fix: drop the applicationId -> instanceIds resolution in SearchController.
Rely on application_id = ? in ClickHouseSearchIndex; keep explicit
instanceIds filtering only when a client passes them.
Related cleanup: the agentIds parameter on StatsStore.statsForRoute /
timeseriesForRoute was silently discarded inside ClickHouseStatsStore, so
per-route stats aggregated across any apps sharing a routeId. Replace with
String applicationId and add application_id to the stats_1m_route filters
so per-route stats are correctly scoped.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The design system doesn't define --text-inverse yet, so SVG icons
on colored badge backgrounds (trace, tap, status) had no fill color.
Define it in index.css as #fff until the DS adds it natively.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
IngestionService.hasAnyTraceData() and ChunkAccumulator only checked
for inputBody/outputBody when setting has_trace_data on executions.
Headers and properties captured via deep tracing were not considered,
causing the trace data indicator to be missing in the exchange list.
DetailService already checked all six fields — this aligns the
ingestion path to match.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Merge the route state update and route catalog upsert blocks to share
one registryService.findById() call instead of two, reducing overhead
on the high-frequency heartbeat path.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add RouteCatalogStore as a third data source in RouteCatalogController so that
/api/v1/routes/catalog surfaces routes with zero executions and routes from
previous app versions that fall within the requested time window.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Wires RouteCatalogStore into CatalogController as a third data source:
routes with zero executions and routes from previous app versions
(within the queried time window) now appear in the unified catalog.
Also clears route_catalog on app dismiss via deleteByApplication().
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Wire RouteCatalogStore into AgentRegistrationController and call upsert
after registration and heartbeat so routes survive server restarts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Routes with zero executions (sub-routes) vanish from the sidebar after
server restart because the catalog is purely in-memory with a ClickHouse
stats fallback that only covers executed routes. This spec describes a
persistent route_catalog table in ClickHouse with lifecycle tracking
(first_seen/last_seen) to reconstruct the sidebar without agent
reconnection and support historical time-window queries.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Use Activity, Cpu, and HeartPulse icons instead of "tps", "cpu", and
"ago" text in compact and expanded app cards. Bump design-system to
v0.1.55 for sidebar footer alignment fix.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-16 15:43:37 +02:00
653 changed files with 91173 additions and 6454 deletions
User-facing data and config endpoints live under `/api/v1/environments/{envSlug}/...`. Env is a path segment, never a query param. The `envSlug` is resolved to an `Environment` bean via the `@EnvPath` argument resolver (`web/EnvironmentPathResolver.java`) — 404 on unknown slug.
**Slugs are immutable after creation** for both environments and apps. Slug regex: `^[a-z0-9][a-z0-9-]{0,63}$`. Validated in `EnvironmentService.create` and `AppService.createApp`. Update endpoints (`PUT`) do not accept a slug field; Jackson drops it as an unknown property.
### Flat-endpoint allow-list
These paths intentionally stay flat (no `/environments/{envSlug}` prefix). Every new endpoint should be env-scoped unless it appears here and the reason is documented.
| Path prefix | Why flat |
|---|---|
| `/api/v1/data/**` | Agent ingestion. JWT `env` claim is authoritative; URL-embedded env would invite spoofing. |
| `/api/v1/diagrams/{contentHash}/render`, `POST /api/v1/diagrams/render` | Content-addressed or stateless. |
| `/api/v1/alerts/notifications/{id}/retry` | Notification IDs are globally unique; no env routing needed. |
| `/api/v1/auth/**` | Pre-auth; no env context exists. |
| `/api/v1/health`, `/prometheus`, `/api-docs/**`, `/swagger-ui/**` | Server metadata. |
## Tenant isolation invariant
ClickHouse is shared across tenants. Every ClickHouse query must filter by `tenant_id` (from `CAMELEER_SERVER_TENANT_ID` env var, resolved via `TenantContext`/config) in addition to `environment`. New controllers added under `/environments/{envSlug}/...` must preserve this — the env filter from the path does not replace the tenant filter.
## User ID conventions
`users.user_id` stores the **bare** identifier:
- Local users: `<username>` (e.g. `admin`, `alice`)
- OIDC users: `oidc:<sub>` (e.g. `oidc:c7a93b…`)
JWT subjects carry a `user:` namespace prefix (`user:admin`, `user:oidc:<sub>`) so `JwtAuthenticationFilter` can distinguish user tokens from agent tokens. All three write paths upsert the **bare** form:
Env-scoped read-path controllers (`AlertController`, `AlertRuleController`, `AlertSilenceController`, `OutboundConnectionAdminController`) strip `"user:"` from `SecurityContextHolder.authentication.name` before using it as an FK. All FKs to `users(user_id)` (e.g. `alert_rules.created_by`, `outbound_connections.created_by`, `alert_reads.user_id`, `user_roles.user_id`, `user_groups.user_id`) therefore reference the bare form. If you add a new controller that needs the acting user id for an FK insert, follow the same strip pattern.
## controller/ — REST endpoints
-`AgentRegistrationController` — POST /register, POST /heartbeat, GET / (list), POST /refresh-token
-`AgentSseController` — GET /sse (Server-Sent Events connection)
-`AgentCommandController` — POST /broadcast, POST /{agentId}, POST /{agentId}/ack
-`AppController` — CRUD /api/v1/apps, POST /{appId}/upload-jar, GET /{appId}/versions
-`DeploymentController` — GET/POST /api/v1/apps/{appId}/deployments, POST /{id}/stop, POST /{id}/promote, GET /{id}/logs
-`EnvironmentAdminController` — CRUD /api/v1/admin/environments, PUT /{id}/jar-retention
-`ExecutionController` — GET /api/v1/executions (search + detail)
-`SearchController` — POST /api/v1/search, GET /routes, GET /top-errors, GET /punchcard
-`LogQueryController` — GET /api/v1/logs (filters: source, application, agentId, exchangeId, level, logger, q, environment, time range)
-`LogIngestionController` — POST /api/v1/data/logs (accepts `List<LogEntry>` JSON array, each entry has `source`: app/agent). Logs WARN for: missing agent identity, unregistered agents, empty payloads, buffer-full drops, deserialization failures. Normal acceptance at DEBUG.
-`CatalogController` — GET /api/v1/catalog (unified app catalog merging PG managed apps + in-memory agents + CH stats), DELETE /api/v1/catalog/{applicationId} (ADMIN: dismiss app, purge all CH data + PG record). Auto-filters discovered apps older than `discoveryttldays` with no live agents.
-`ChunkIngestionController` — POST /api/v1/ingestion/chunk/{executions|metrics|diagrams}
-`UserAdminController` — CRUD /api/v1/admin/users, POST /{id}/roles, POST /{id}/set-password
-`OidcConfigAdminController`— GET/POST /api/v1/admin/oidc, POST /test
-`SensitiveKeysAdminController` — GET/PUT /api/v1/admin/sensitive-keys. GET returns 200 with config or 204 if not configured. PUT accepts `{ keys: [...] }` with optional `?pushToAgents=true` to fan out merged keys to all LIVE agents. Stored in `server_config` table (key `sensitive_keys`).
-`AuditLogController` — GET /api/v1/admin/audit
-`MetricsController` — GET /api/v1/metrics, GET /timeseries
-`DiagramController` — GET /api/v1/diagrams/{id}, POST /
-`DiagramRenderController` — POST /api/v1/diagrams/render (ELK layout)
-`ClaimMappingAdminController` — CRUD /api/v1/admin/claim-mappings, POST /test (accepts inline rules + claims for preview without saving)
-`UsageAnalyticsController` — GET /api/v1/admin/usage (ClickHouse usage_events)
### Env-scoped (user-facing data & config)
-`AppController` — `/api/v1/environments/{envSlug}/apps`. GET list / POST create / GET `{appSlug}` / DELETE `{appSlug}` / GET `{appSlug}/versions` / POST `{appSlug}/versions` (JAR upload) / PUT `{appSlug}/container-config` / GET `{appSlug}/dirty-state` (returns `DirtyStateResponse{dirty, lastSuccessfulDeploymentId, differences}` — compares current JAR+config against last RUNNING deployment snapshot; dirty=true when no snapshot exists). App slug uniqueness is per-env (`(env, app_slug)` is the natural key). `CreateAppRequest` body has no env (path), validates slug regex. Injects `DirtyStateCalculator` bean (registered in `RuntimeBeanConfig`, requires `ObjectMapper` with `JavaTimeModule`).
-`DeploymentController` — `/api/v1/environments/{envSlug}/apps/{appSlug}/deployments`. GET list / POST create (body `{ appVersionId }`) / POST `{id}/stop` / POST `{id}/promote` (body `{ targetEnvironment: slug }` — target app slug must exist in target env) / GET `{id}/logs`. All lifecycle ops (`POST /` deploy, `POST /{id}/stop`, `POST /{id}/promote`) audited under `AuditCategory.DEPLOYMENT`. Action codes: `deploy_app`, `stop_deployment`, `promote_deployment`. Acting user resolved via the `user:` prefix-strip convention; both SUCCESS and FAILURE branches write audit rows. `created_by` (TEXT, nullable) populated from `SecurityContextHolder` and surfaced on the `Deployment` DTO.
-`ApplicationConfigController` — `/api/v1/environments/{envSlug}`. GET `/config` (list), GET/PUT `/apps/{appSlug}/config`, GET `/apps/{appSlug}/processor-routes`, POST `/apps/{appSlug}/config/test-expression`. PUT accepts `?apply=staged|live` (default `live`). `live` saves to DB and pushes `CONFIG_UPDATE` SSE to live agents in this env (existing behavior); `staged` saves to DB only, skipping the SSE push — used by the unified app deployment page. Audit action is `stage_app_config` for staged writes, `update_app_config` for live. Invalid `apply` values return 400.
-`AppSettingsController` — `/api/v1/environments/{envSlug}`. GET `/app-settings` (list), GET/PUT/DELETE `/apps/{appSlug}/settings`. ADMIN/OPERATOR only.
-`SearchController` — `/api/v1/environments/{envSlug}`. GET `/executions`, POST `/executions/search`, GET `/stats`, `/stats/timeseries`, `/stats/timeseries/by-app`, `/stats/timeseries/by-route`, `/stats/punchcard`, `/attributes/keys`, `/errors/top`. GET `/executions` accepts repeat `attr` query params: `attr=order` (key-exists), `attr=order:47` (exact), `attr=order:4*` (wildcard — `*` maps to SQL LIKE `%`). First `:` splits key/value; later colons stay in the value. Invalid keys → 400. POST `/executions/search` accepts the same filters via `SearchRequest.attributeFilters` in the body.
-`LogQueryController` — GET `/api/v1/environments/{envSlug}/logs` (filters: source (multi, comma-split, OR-joined), level (multi, comma-split, OR-joined), application, agentId, exchangeId, logger, q, time range, instanceIds (multi, comma-split, AND-joined as WHERE instance_id IN (...) — used by the Checkpoint detail drawer to scope logs to a deployment's replicas); sort asc/desc). Cursor-paginated, returns `{ data, nextCursor, hasMore, levelCounts }`; cursor is base64url of `"{timestampIso}|{insert_id_uuid}"` — same-millisecond tiebreak via the `insert_id` UUID column on `logs`.
-`RouteCatalogController` — GET `/api/v1/environments/{envSlug}/routes` (merged route catalog from registry + ClickHouse; env filter unconditional).
-`RouteMetricsController` — GET `/api/v1/environments/{envSlug}/routes/metrics`, GET `/api/v1/environments/{envSlug}/routes/metrics/processors`.
-`AgentListController` — GET `/api/v1/environments/{envSlug}/agents` (registered agents with runtime metrics, filtered to env).
-`AgentEventsController` — GET `/api/v1/environments/{envSlug}/agents/events` (lifecycle events; cursor-paginated, returns `{ data, nextCursor, hasMore }`; order `(timestamp DESC, insert_id DESC)`; cursor is base64url of `"{timestampIso}|{insert_id_uuid}"` — `insert_id` is a stable UUID column used as a same-millisecond tiebreak).
-`AgentMetricsController` — GET `/api/v1/environments/{envSlug}/agents/{agentId}/metrics` (JVM/Camel metrics). Rejects cross-env agents (404) as defence-in-depth.
-`DiagramRenderController` — GET `/api/v1/environments/{envSlug}/apps/{appSlug}/routes/{routeId}/diagram` returns the most recent diagram for (app, env, route) via `DiagramStore.findLatestContentHashForAppRoute`. Registry-independent — routes whose publishing agents were removed still resolve. Also GET `/api/v1/diagrams/{contentHash}/render` (flat — content hashes are globally unique), the point-in-time path consumed by the exchange viewer via `ExecutionDetail.diagramContentHash`.
-`AlertRuleController` — `/api/v1/environments/{envSlug}/alerts/rules`. GET list / POST create / GET `{id}` / PUT `{id}` / DELETE `{id}` / POST `{id}/enable` / POST `{id}/disable` / POST `{id}/render-preview` / POST `{id}/test-evaluate`. OPERATOR+ for mutations, VIEWER+ for reads. CRITICAL: attribute keys in `ExchangeMatchCondition.filter.attributes` are validated at rule-save time against `^[a-zA-Z0-9._-]+$` — they are later inlined into ClickHouse SQL. `AgentLifecycleCondition` is allowlist-only — the `AgentLifecycleEventType` enum (REGISTERED / RE_REGISTERED / DEREGISTERED / WENT_STALE / WENT_DEAD / RECOVERED) plus the record compact ctor (non-empty `eventTypes`, `withinSeconds ≥ 1`) do the validation; custom agent-emitted event types are tracked in backlog issue #145. Webhook validation: verifies `outboundConnectionId` exists and `isAllowedInEnvironment`. Null notification templates default to `""` (NOT NULL constraint). Audit: `ALERT_RULE_CHANGE`.
-`AlertController` — `/api/v1/environments/{envSlug}/alerts`. GET list (inbox filtered by userId/groupIds/roleNames via `InAppInboxQuery`; optional multi-value `state`, `severity`, tri-state `acked`, tri-state `read` query params; soft-deleted rows always excluded) / GET `/unread-count`/ GET`{id}` / POST `{id}/ack` / POST `{id}/read` / POST `/bulk-read` / POST `/bulk-ack` (VIEWER+) / DELETE `{id}` (OPERATOR+, soft-delete) / POST `/bulk-delete` (OPERATOR+) / POST `{id}/restore` (OPERATOR+, clears `deleted_at`). `requireLiveInstance` helper returns 404 on soft-deleted rows; `restore` explicitly fetches regardless of `deleted_at`. `BulkIdsRequest` is the shared body for bulk-read/ack/delete (`{ instanceIds }`). `AlertDto` includes `readAt`; `deletedAt` is intentionally NOT on the wire. Inbox SQL: `? = ANY(target_user_ids) OR target_group_ids && ? OR target_role_names && ?` — requires at least one matching target (no broadcast concept).
-`AlertSilenceController` — `/api/v1/environments/{envSlug}/alerts/silences`. GET list / POST create / DELETE `{id}`. 422 if `endsAt <= startsAt`. OPERATOR+ for mutations, VIEWER+ for list. Audit: `ALERT_SILENCE_CHANGE`.
-`AlertNotificationController` — Dual-path (no class-level prefix). GET `/api/v1/environments/{envSlug}/alerts/{alertId}/notifications` (VIEWER+); POST `/api/v1/alerts/notifications/{id}/retry` (OPERATOR+, flat — notification IDs globally unique). Retry resets attempts to 0 and sets `nextAttemptAt = now`.
### Env admin (env-slug-parameterized, not env-scoped data)
-`EnvironmentAdminController` — `/api/v1/admin/environments`. GET list / POST create / GET `{envSlug}` / PUT `{envSlug}` / DELETE `{envSlug}` / PUT `{envSlug}/default-container-config` / PUT `{envSlug}/jar-retention`. Slug immutable — PUT body has no slug field; any slug supplied is dropped by Jackson. Slug validated on POST. `UpdateEnvironmentRequest` carries `color` (nullable); unknown values rejected with 400 via `EnvironmentColor.isValid`. Null/absent color preserves the existing value.
-`AgentRegistrationController` — POST `/register` (requires `environmentId` in body; 400 if missing), POST `/{id}/refresh` (rejects tokens with no `env` claim), POST `/{id}/heartbeat` (env from body preferred, JWT fallback; 400 if neither), POST `/{id}/deregister`.
-`AgentCommandController` — POST `/{agentId}/commands`, POST `/groups/{group}/commands`, POST `/commands` (broadcast), POST `/{agentId}/commands/{commandId}/ack`, POST `/{agentId}/replay`.
-`AgentConfigController` — GET `/api/v1/agents/config`. Agent-authoritative config read: resolves (app, env) from JWT subject → registry (registry miss falls back to JWT env claim; no registry entry → 404 since application can't be derived).
### Ingestion (agent-only, JWT-authoritative)
-`LogIngestionController` — POST `/api/v1/data/logs` (accepts `List<LogEntry>`; WARNs on missing identity, unregistered agents, empty payloads, buffer-full drops).
-`EventIngestionController` — POST `/api/v1/data/events`.
-`ChunkIngestionController` — POST `/api/v1/data/executions`. Accepts a single `ExecutionChunk` or an array (fields include `exchangeId`, `applicationId`, `instanceId`, `routeId`, `status`, `startTime`, `endTime`, `durationMs`, `chunkSeq`, `final`, `processors: FlatProcessorRecord[]`). The accumulator merges non-final chunks by exchangeId and emits the merged envelope on the final chunk or on stale timeout. Legacy `ExecutionController` / `RouteExecution` shape is retired.
-`MetricsController` — POST `/api/v1/data/metrics`.
-`DiagramController` — POST `/api/v1/data/diagrams` (resolves applicationId + environment from the agent registry keyed on JWT subject; stamps both on the stored `TaggedDiagram`).
### Cross-env discovery (flat)
-`CatalogController` — GET `/api/v1/catalog` (merges managed apps + in-memory agents + CH stats; optional `?environment=` filter). DELETE `/api/v1/catalog/{applicationId}` (ADMIN: dismiss app, purge all CH data + PG record).
### Admin (cross-env, flat)
-`UserAdminController` — CRUD `/api/v1/admin/users`, POST `/{id}/roles`, POST `/{id}/set-password`.
-`OidcConfigAdminController` — GET/POST `/api/v1/admin/oidc`, POST `/test`.
-`OutboundConnectionAdminController` — `/api/v1/admin/outbound-connections`. GET list / POST create / GET `{id}` / PUT `{id}` / DELETE `{id}` / POST `{id}/test` / GET `{id}/usage`. RBAC: list/get/usage ADMIN|OPERATOR; mutations + test ADMIN.
-`SensitiveKeysAdminController` — GET/PUT `/api/v1/admin/sensitive-keys`. GET returns 200 or 204 if not configured. PUT accepts `{ keys: [...] }` with optional `?pushToAgents=true`. Fan-out iterates every distinct `(application, environment)` slice — intentional global baseline + per-env overrides.
-`ClaimMappingAdminController` — CRUD `/api/v1/admin/claim-mappings`, POST `/test`.
-`LicenseAdminController` — GET/POST `/api/v1/admin/license`. ADMIN only. GET returns `{state, invalidReason, envelope, lastValidatedAt?}` — the raw token is deliberately omitted; only the parsed `LicenseInfo` envelope is exposed. POST delegates to `LicenseService.install(token, userId, "api")` (acting userId resolved via the `user:` prefix-strip convention) — install/replace/reject all flow through `LicenseService` so audit, persistence, and `LicenseChangedEvent` publishing are uniform.
-`LicenseUsageController` — GET `/api/v1/admin/license/usage`. Returns license `state`, `expiresAt`/`daysRemaining`/`gracePeriodDays`/`tenantId`/`label`/`lastValidatedAt`, the `LicenseMessageRenderer.forState(...)` message, and a `limits[]` array (`{key, current, cap, source}`) covering every effective-limits key. `source` is `"license"` when the cap came from the license override map, `"default"` otherwise. `max_agents` reads from `AgentRegistryService.liveCount()`; all other counts come from `LicenseUsageReader.snapshot()`.
-`AuditLogController` — GET `/api/v1/admin/audit`.
-`RbacStatsController` — GET `/api/v1/admin/rbac/stats`.
-`UsageAnalyticsController` — GET `/api/v1/admin/usage` (ClickHouse `usage_events`).
-`ClickHouseAdminController` — GET `/api/v1/admin/clickhouse/**` (conditional on `infrastructureendpoints` flag).
-`DatabaseAdminController` — GET `/api/v1/admin/database/**` (conditional on `infrastructureendpoints` flag).
-`ServerMetricsAdminController` — `/api/v1/admin/server-metrics/**`. GET `/catalog`, GET `/instances`, POST `/query`. Generic read API over the `server_metrics` ClickHouse table so SaaS dashboards don't need direct CH access. Delegates to `ServerMetricsQueryStore` (impl `ClickHouseServerMetricsQueryStore`). Visibility matches ClickHouse/Database admin: `@ConditionalOnProperty(infrastructureendpoints, matchIfMissing=true)` + class-level `@PreAuthorize("hasRole('ADMIN')")`. Validation: metric/tag regex `^[a-zA-Z0-9._]+$`, statistic regex `^[a-z_]+$`, `to - from ≤ 31 days`, stepSeconds ∈ [10, 3600], response capped at 500 series. `IllegalArgumentException` → 400. `/query` supports `raw` + `delta` modes (delta does per-`server_instance_id` positive-clipped differences, then aggregates across instances). Derived `statistic=mean` for timers computes `sum(total|total_time)/sum(count)` per bucket.
### Auth (flat)
-`UiAuthController` — `/api/v1/auth` (login, refresh, me). Local username/password against env-var admin or DB BCrypt hash. Lockout after 5 failed attempts.
-`OidcAuthController` — `/api/v1/auth/oidc` (config, callback). Code → token exchange. Roles via custom JWT claim, claim mapping rules, or default roles.
-`AuthCapabilitiesController` — `GET /api/v1/auth/capabilities` (unauthenticated). Reports `{oidc:{enabled, providerName, primary}, localAccounts:{enabled, adminRecoveryOnly}}` so the SPA renders the login page deterministically. `oidc.primary == oidc.enabled`; `localAccounts.adminRecoveryOnly == oidc.primary`. `providerName` is best-effort label via `OidcProviderNameDeriver` (Logto / Keycloak / Auth0 / Okta / Single Sign-On). The SPA hides the local form behind `?local` when `adminRecoveryOnly` is true.
### Other (flat)
-`DetailController` — GET `/api/v1/executions/{executionId}` + processor snapshot endpoints.
-`MetricsController` — exposes `/api/v1/metrics` and `/api/v1/prometheus` (server-side Prometheus scrape endpoint).
-`DeploymentExecutor` — @Async staged deploy: PRE_FLIGHT -> PULL_IMAGE -> CREATE_NETWORK -> START_REPLICAS -> HEALTH_CHECK -> SWAP_TRAFFIC -> COMPLETE. Container names are `{tenantId}-{envSlug}-{appSlug}-{replicaIndex}-{generation}`, where `generation` is the first 8 chars of the deployment UUID — old and new replicas coexist during a blue/green swap. Per-replica `CAMELEER_AGENT_INSTANCEID` env var is`{envSlug}-{appSlug}-{replicaIndex}-{generation}`. Branches on `DeploymentStrategy.fromWire(config.deploymentStrategy())`: **blue-green** (default) starts all N → waits for all healthy → stops old (partial health = FAILED, preserves old untouched); **rolling** replaces replicas one at a time with rollback only for in-flight new containers (already-replaced old stay stopped; un-replaced old keep serving). DEGRADED is now only set by `DockerEventMonitor` post-deploy, never by the executor. **License compute caps**: at PRE_FLIGHT (after `ConfigMerger.resolve`, before image pull / container creation) the executor consults `LicenseUsageReader.computeUsage()` (PG aggregate over non-stopped deployments) and runs three `LicenseEnforcer.assertWithinCap(...)` checks for `max_total_cpu_millis`, `max_total_memory_mb`, and `max_total_replicas`. A `LicenseCapExceededException` propagates to the surrounding `try/catch` which marks the deployment FAILED with the cap message in `deployments.error_message`.
-`TraefikLabelBuilder` — generates Traefik Docker labels for path-based or subdomain routing. Also emits `cameleer.replica` and `cameleer.instance-id` labels per container for labels-first identity.
-`TraefikLabelBuilder` — generates Traefik Docker labels for path-based or subdomain routing. Per-container identity labels: `cameleer.replica` (index), `cameleer.generation` (deployment-scoped 8-char id — for Prometheus/Grafana deploy-boundary annotations), `cameleer.instance-id` (`{envSlug}-{appSlug}-{replicaIndex}-{generation}`). Router/service label keys are generation-agnostic so load balancing spans old + new replicas during a blue/green overlap.
-`PrometheusLabelBuilder` — generates Prometheus Docker labels (`prometheus.scrape/path/port`) per runtime type for `docker_sd_configs` auto-discovery
-`ContainerLogForwarder` — streams Docker container stdout/stderr to ClickHouse with `source='container'`. One follow-stream thread per container, batches lines every 2s/50 lines via `ClickHouseLogStore.insertBufferedBatch()`. 60-second max capture timeout.
-`DisabledRuntimeOrchestrator` — no-op when runtime not enabled
@@ -60,14 +137,16 @@ paths:
## metrics/ — Prometheus observability
-`ServerMetrics` — centralized business metrics: gauges (agents by state, SSE connections, buffer depths), counters (ingestion drops, agent transitions, deployment outcomes, auth failures), timers (flush duration, deployment duration). Exposed via `/api/v1/prometheus`.
-`ServerInstanceIdConfig` — `@Configuration`, exposes `@Bean("serverInstanceId") String`. Resolution precedence: `cameleer.server.instance-id` property → `HOSTNAME` env → `InetAddress.getLocalHost()` → random UUID. Fixed at boot; rotates across restarts so counters restart cleanly.
-`ServerMetricsSnapshotScheduler` — `@Scheduled(fixedDelayString = "${cameleer.server.self-metrics.interval-ms:60000}")`. Walks `MeterRegistry.getMeters()` each tick, emits one `ServerMetricSample` per `Measurement` (Timer/DistributionSummary produce multiple rows per meter — one per Micrometer `Statistic`). Skips non-finite values; logs and swallows store failures. Disabled via `cameleer.server.self-metrics.enabled=false` (`@ConditionalOnProperty`). Write-only — no query endpoint yet; inspect via `/api/v1/admin/clickhouse/query`.
-`PostgresDeploymentRepository` — includes JSONB replica_states, deploy_stage, findByContainerId
-`PostgresDeploymentRepository` — includes JSONB replica_states, deploy_stage, findByContainerId. Also carries `deployed_config_snapshot` JSONB (Flyway V3) populated by `DeploymentExecutor` via `saveDeployedConfigSnapshot(UUID, DeploymentConfigSnapshot)` on successful RUNNING transition. Consumed by `DirtyStateCalculator` for the `/apps/{slug}/dirty-state` endpoint and by the UI for checkpoint restore.
-`PostgresAppSettingsRepository`, `PostgresApplicationConfigRepository`, `PostgresThresholdRepository`. Both `app_settings` and `application_config` are env-scoped (PK `(app_id, environment)` / `(application, environment)`); finders take `(app, env)` — no env-agnostic variants.
-`ClickHouseUsageTracker` — usage_events for billing
-`ClickHouseRouteCatalogStore` — persistent route catalog with first_seen cache, warm-loaded on startup
-`ClickHouseServerMetricsStore` — periodic dumps of the server's own Micrometer registry into the `server_metrics` table. Tenant-stamped (bound at the scheduler, not the bean); no `environment` column (server straddles envs). Batch-insert via `JdbcTemplate.batchUpdate` with `Map(String, String)` tag binding. Written by `ServerMetricsSnapshotScheduler`.
-`ClickHouseServerMetricsQueryStore` — read side of `server_metrics` for dashboards. Implements `ServerMetricsQueryStore`. `catalog(from,to)` returns name+type+statistics+tagKeys, `listInstances(from,to)` returns server_instance_ids with first/last seen, `query(request)` builds bucketed time-series with `raw` or `delta` mode and supports a derived `mean` statistic for timers. All identifier inputs regex-validated; tenant_id always bound; max range 31 days; series count capped at 500. Exposed via `ServerMetricsAdminController`.
-`AlertEvaluatorJob` — @Scheduled tick driver; per-rule claim/release via `AlertRuleRepository`, dispatches to per-kind `ConditionEvaluator`, persists advanced cursor on release via `AlertRule.withEvalState`.
-`BatchResultApplier` — `@Component` that wraps a single rule's tick outcome (`EvalResult.Batch` = `firings` + `nextEvalState`) in one `@Transactional` boundary: instance upserts + notification enqueues + cursor advance commit atomically or roll back together. This is the exactly-once-per-exchange guarantee for `PER_EXCHANGE` fire mode.
-`crypto/SecretCipher` — AES-GCM symmetric cipher with key derived via HMAC-SHA256(jwtSecret, "cameleer-outbound-secret-v1"). Ciphertext format: base64(IV(12 bytes) || GCM output with 128-bit tag). `encrypt` throws `IllegalStateException`; `decrypt` throws `IllegalArgumentException` on tamper/wrong-key/malformed.
-`storage/PostgresOutboundConnectionRepository` — JdbcTemplate impl. `save()` upserts by id; JSONB serialization via ObjectMapper; UUID arrays via `ConnectionCallback`. Reads `created_by`/`updated_by` as String (= users.user_id TEXT).
-`OutboundConnectionServiceImpl` — service layer. Tenant bound at construction via `cameleer.server.tenant.id` property. Uniqueness check via `findByName`. Narrowing-envs guard: rejects update that removes envs while rules reference the connection (rulesReferencing stubbed in Plan 01, wired in Plan 02). Delete guard: rejects if referenced by rules.
-`controller/OutboundConnectionAdminController` — REST controller. Class-level `@PreAuthorize("hasRole('ADMIN')")` defaults; GETs relaxed to ADMIN|OPERATOR. Resolves acting user id via the user-id convention (strip `"user:"` from `authentication.name` → matches `users.user_id` FK). Audit via `AuditCategory.OUTBOUND_CONNECTION_CHANGE`.
-`dto/OutboundConnectionRequest` — Bean Validation: `@NotBlank` name, `@Pattern("^https://.+")` url, `@NotNull` method/tlsTrustMode/auth. Compact ctor throws `IllegalArgumentException` if TRUST_PATHS with empty paths list.
-`dto/OutboundConnectionDto` — response DTO. `hmacSecretSet: boolean` instead of the ciphertext; `authKind: OutboundAuthKind` instead of the full auth config.
-`dto/OutboundConnectionTestResult` — result of POST `/{id}/test`: status, latencyMs, responseSnippet (first 512 chars), tlsProtocol/cipherSuite/peerCertSubject (protocol is "TLS" stub; enriched in Plan 02 follow-up), error (nullable).
-`LicenseService` — install / replace / revalidate mediator. `install(token, installedBy, source)` validates via `LicenseValidator`, on failure marks the gate INVALID + audits `reject_license` + publishes `LicenseChangedEvent` and rethrows; on success persists via `LicenseRepository.upsert(...)`, mutates `LicenseGate`, audits `install_license` or `replace_license` (detects existing row), and publishes `LicenseChangedEvent`. `loadInitial(envToken, fileToken)` boot precedence env > file > DB; ABSENT publishes a `LicenseChangedEvent(ABSENT, null)`. `revalidate()` re-runs validation against the persisted token, on success bumps `last_validated_at`; on failure marks INVALID and audits `revalidate_license` FAILURE. `getTenantId()` exposes the tenant for downstream lookups.
-`PostgresLicenseRepository` — JdbcTemplate impl of `LicenseRepository`. Targets PG `license` table (V5). Upsert via `INSERT ... ON CONFLICT (tenant_id) DO UPDATE`.
-`LicenseChangedEvent` — Spring application event: `(LicenseState state, LicenseInfo current)`. Published on every install / replace / revalidate / boot-time ABSENT path so downstream listeners (retention policy, metrics, etc.) react uniformly.
-`LicenseEnforcer` — `@Component`. `assertWithinCap(String limitKey, long currentUsage, long requestedDelta)` consults `LicenseGate.getEffectiveLimits()`. On overflow increments `cameleer_license_cap_rejections_total{limit=...}`, emits an `AuditCategory.LICENSE / cap_exceeded` audit row when `AuditService` is wired (try/catch + log.warn so audit-write failures don't suppress the 403), and throws `LicenseCapExceededException`. Unknown limit keys propagate `IllegalArgumentException` from `LicenseLimits.get(...)` (programmer error, not a 403).
-`LicenseUsageReader` — `@Component` over PG. `snapshot()` returns a `Map<String,Long>` of (max_environments, max_apps, max_users, max_outbound_connections, max_alert_rules, max_total_cpu_millis, max_total_memory_mb, max_total_replicas) from PG row counts and a SUM over non-stopped deployments' `deployed_config_snapshot.containerConfig` (replicas × cpuLimit / memoryLimitMb). `computeUsage()` returns the typed `ComputeUsage(cpuMillis, memoryMb, replicas)` tuple consumed by `DeploymentExecutor` PRE_FLIGHT cap checks. `agentCount(int)` echoes a registry-supplied live count (registry is in-memory; not stored in PG).
-`LicenseCapExceededException` — typed `RuntimeException(limitKey, current, cap)` with accessors. Mapped to HTTP 403 by `LicenseExceptionAdvice`.
-`LicenseExceptionAdvice` — `@ControllerAdvice` mapping `LicenseCapExceededException` → 403 with body `{error:"license cap reached", limit, current, cap, state, message}` where `message` is `LicenseMessageRenderer.forCap(state, info, limit, current, cap, invalidReason)`.
-`LicenseMessageRenderer` — pure formatter (utility class, no DI). `forCap(state, info, limit, current, cap[, invalidReason])` per-state human messages for cap-rejection responses; `forState(state, info[, invalidReason])` shorter state-only messages for the `/usage` endpoint and metrics surfaces.
-`RetentionPolicyApplier` — `@EventListener(LicenseChangedEvent.class) @Async`. For each environment × table in the static `SPECS` list (`executions`, `processor_executions`, `logs`, `agent_metrics`, `agent_events`) computes `effective = min(licenseCap, env.configuredRetentionDays)` and emits `ALTER TABLE <t> MODIFY TTL toDateTime(<col>) + INTERVAL <n> DAY DELETE WHERE environment = '<slug>'`. ClickHouse failures are logged and swallowed (best-effort; never propagates to the originating license install/revalidate). `route_diagrams` (no TTL clause) and `server_metrics` (no environment column) are intentionally excluded.
-`LicenseRevalidationJob` — `@Component`. `@Scheduled(cron = "0 0 3 * * *")` daily revalidation; `@EventListener(ApplicationReadyEvent.class) @Async` 60-second post-startup tick to catch ABSENT→ACTIVE when a license was inserted between server starts. Both paths call `LicenseService.revalidate()` and swallow scheduler-thread crashes.
-`LicenseMetrics` — `@Component`. Registers Micrometer gauges: `cameleer_license_state{state=...}` (one-hot per `LicenseState`), `cameleer_license_days_remaining` (negative when ABSENT/INVALID), `cameleer_license_last_validated_age_seconds` (0 when no DB row). Refreshed eagerly on `LicenseChangedEvent` via `@EventListener` and lazily every 60s via `@Scheduled(fixedDelay = 60_000)`.
-`RuntimeBeanConfig` — DeploymentExecutor, AppService, EnvironmentService. Wires `CreateGuard` instances per service from `LicenseEnforcer.assertWithinCap(...)` so creation paths (Environment, App, Agent) consult license caps without core depending on the app module.
-`LicenseBeanConfig` — license bean topology in dependency order: `LicenseGate` → `LicenseValidator` (when `cameleer.server.license.publickey` is unset, an always-failing override is returned so any loaded token still routes through `install()` and is audited as INVALID, never silently dropped) → `LicenseService` → `LicenseBootLoader` (`@PostConstruct` drives `loadInitial(envToken, fileToken)` once the context is ready; resolution order env var > license file > persisted DB row).
- CI workflow: `.gitea/workflows/ci.yml` — build -> docker -> deploy on push to main or feature branches
- CI workflow: `.gitea/workflows/ci.yml` — build -> docker -> deploy on push to main or feature branches. `paths-ignore` skips the whole pipeline for docs-only / `.planning/` / `.claude/` / `*.md` changes (push and PR triggers).
- Build caches (parallel `actions/cache@v4` steps in the `build` job): `~/.m2/repository` (key on all `pom.xml`), `~/.npm` (key on `ui/package-lock.json`), `ui/node_modules/.vite` (key on `ui/package-lock.json` + `ui/vite.config.ts`). UI install uses `npm ci --prefer-offline --no-audit --fund=false` so the npm cache is the primary source.
- Maven build performance (set in `pom.xml` and `cameleer-server-app/pom.xml`): `useIncrementalCompilation=true` on the compiler plugin; Surefire uses `forkCount=1C` + `reuseForks=true` (one JVM per CPU core, reused across test classes); Failsafe keeps `forkCount=1` + `reuseForks=true`. Unit tests must not rely on per-class JVM isolation.
- UI build script (`ui/package.json`): `build` is `vite build` only — the type-check pass was split out into `npm run typecheck` (run separately when you want a full `tsc --noEmit` sweep).
- Docker: multi-stage build (`Dockerfile`), `$BUILDPLATFORM` for native Maven on ARM64 runner, amd64 runtime. `docker-entrypoint.sh` imports `/certs/ca.pem` into JVM truststore before starting the app (supports custom CAs for OIDC discovery without `CAMELEER_SERVER_SECURITY_OIDCTLSSKIPVERIFY`).
-`REGISTRY_TOKEN` build arg required for `cameleer-common` dependency resolution
-`Environment` — record: id, slug, displayName, production, enabled, defaultContainerConfig, jarRetentionCount, color, createdAt, executionRetentionDays, logRetentionDays, metricRetentionDays. `color` is one of the 8 preset palette values validated by `EnvironmentColor.VALUES` and CHECK-constrained in PostgreSQL (V2 migration). The 3 retention day fields (V5) are `int`-typed (not nullable, since unlimited has no use-case), default to 1 day per the V5 `NOT NULL DEFAULT 1`, validated >= 1 in the canonical constructor.
-`DeploymentStatus` — enum: STOPPED, STARTING, RUNNING, DEGRADED, STOPPING, FAILED. `DEGRADED` is reserved for post-deploy drift (a replica died after RUNNING); `DeploymentExecutor` now marks partial-healthy deploys FAILED, not DEGRADED.
-`DeploymentStrategy` — enum: BLUE_GREEN, ROLLING. Stored on `ResolvedContainerConfig.deploymentStrategy` as kebab-case string (`"blue-green"` / `"rolling"`). `fromWire(String)` is the only conversion entry point; unknown/null inputs fall back to BLUE_GREEN so the executor dispatch site never null-checks or throws.
-`DeploymentService` — createDeployment (calls `deleteFailedByAppAndEnvironment` first so FAILED rows don't pile up; STOPPED rows are preserved as restorable checkpoints), markRunning, markFailed, markStopped
-`ResolvedContainerConfig` — record: typed config with memoryLimitMb, memoryReserveMb, cpuRequest, cpuLimit, appPort, exposedPorts, customEnvVars, stripPathPrefix, sslOffloading, routingMode, routingDomain, serverUrl, replicas, deploymentStrategy, routeControlEnabled, replayEnabled, runtimeType, customArgs, extraNetworks, externalRouting (default `true`; when `false`, `TraefikLabelBuilder` strips all `traefik.*` labels so the container is not publicly routed), certResolver (server-wide, sourced from `CAMELEER_SERVER_RUNTIME_CERTRESOLVER`; when blank the `tls.certresolver` label is omitted — use for dev installs with a static TLS store)
-`RoutingMode` — enum for routing strategies
-`ConfigMerger` — pure function: resolve(globalDefaults, envConfig, appConfig) -> ResolvedContainerConfig
-`CreateGuard` — `@FunctionalInterface`. `void check(long current)` — implementations throw to abort creation. `NOOP` constant is the default. Consulted by `EnvironmentService.create`, `AppService.createApp`, and `AgentRegistryService.register` so license caps can be enforced from the app module without leaking Spring or app-only types into core. Wired in `LicenseBeanConfig` to a `LicenseEnforcer.assertWithinCap(...)` call per limit key.
The pure license **contract types** live in the separate `cameleer-license-api` module under package `com.cameleer.license` (no Spring, no server-runtime deps) so consumers like `cameleer-license-minter` and `cameleer-saas` can use them without inheriting server internals. Server-core only contains the runtime state holder (`LicenseGate`).
Contract types in `cameleer-license-api` (package `com.cameleer.license`):
-`LicenseInfo` — record: `(UUID licenseId, String tenantId, String label, Map<String,Integer> limits, Instant issuedAt, Instant expiresAt, int gracePeriodDays)`. `isExpired()` true once `now > expiresAt + gracePeriodDays`; `isAfterRawExpiry()` true once `now > expiresAt`. Constructed via `LicenseValidator`; canonical ctor null-checks all required fields and rejects blank tenantId / negative grace.
-`LicenseLimits` — typed limits container backed by `Map<String,Integer>`. `defaultsOnly()` returns the `DefaultTierLimits.DEFAULTS` view; `mergeOverDefaults(overrides)` produces the license-overrides UNION default tier. `get(String key)` returns the cap; throws `IllegalArgumentException` for unknown keys (programmer error). `isDefaultSourced(key, license)` reports whether a key fell through to the default tier.
-`DefaultTierLimits` — immutable `LinkedHashMap` of constants for the no-license fallback tier: `max_environments=1, max_apps=3, max_agents=5, max_users=3, max_outbound_connections=1, max_alert_rules=2, max_total_cpu_millis=2000, max_total_memory_mb=2048, max_total_replicas=5, max_execution_retention_days=1, max_log_retention_days=1, max_metric_retention_days=1, max_jar_retention_count=3`.
-`LicenseValidator` — verifies signed token. Constructor `(String publicKeyBase64, String expectedTenantId)` decodes an X.509 Ed25519 public key. `validate(String token)` splits `payload.signature`, verifies the Ed25519 signature, parses the JSON payload, enforces `tenantId == expectedTenantId`, and returns `LicenseInfo`. Throws `SecurityException` on signature mismatch / `IllegalArgumentException` on parse failure / expired payload.
-`LicenseStateMachine` — pure classifier. `classify(LicenseInfo, String invalidReason)` returns `INVALID` if a reason is set, `ABSENT` if no license, `ACTIVE` if `now <= expiresAt`, `GRACE` if expired but within grace window, `EXPIRED` otherwise.
Runtime state holder in server-core (package `com.cameleer.server.core.license`):
-`LicenseGate` — runtime state holder (thread-safe via `AtomicReference<Snapshot>`). `getCurrent()` returns the current `LicenseInfo` (null when ABSENT/INVALID); `getState()` delegates to `LicenseStateMachine.classify(...)`; `getEffectiveLimits()` returns license-overrides UNION defaults in `ACTIVE`/`GRACE`, defaults-only otherwise. `getInvalidReason()`, `load(LicenseInfo)`, `markInvalid(String reason)`, `clear()` are the mutators. `getLimit(key, defaultValue)` shorthand swallows unknown-key errors.
-`SearchService` — search, count, stats, statsForApp, statsForRoute, timeseries, timeseriesForApp, timeseriesForRoute, timeseriesGroupedByApp, timeseriesGroupedByRoute, slaCompliance, slaCountsByApp, slaCountsByRoute, topErrors, activeErrorTypes, punchcard, distinctAttributeKeys. `statsForRoute`/`timeseriesForRoute` take `(routeId, applicationId)` — app filter is applied to `stats_1m_route`.
-`SearchRequest` / `SearchResult` — search DTOs. `SearchRequest.attributeFilters: List<AttributeFilter>` carries structured facet filters for execution attributes — key-only (exists), exact (key=value), or wildcard (`*` in value). The 21-arg legacy ctor is preserved for call-site churn; the compact ctor normalises null → `List.of()`.
-`AttributeFilter(key, value)` — record with key regex `^[a-zA-Z0-9._-]+$` (inlined into SQL, same constraint as alerting), `value == null` means key-exists, `value` containing `*` becomes a SQL LIKE pattern via `toLikePattern()`.
-`ExecutionStats`, `ExecutionSummary` — stats aggregation records
-`StatsTimeseries`, `TopError` — timeseries and error DTOs
-`ExecutionStore`, `MetricsStore`, `MetricsQueryStore`, `StatsStore`, `DiagramStore`,`RouteCatalogStore`,`SearchIndex`, `LogIndex` — interfaces. `DiagramStore.findLatestContentHashForAppRoute(appId, routeId, env)` resolves the latest diagram by (app, env, route) without consulting the agent registry, so routes whose publishing agents were removed between app versions still resolve. `findContentHashForRoute(route, instance)` is retained for the ingestion path that stamps a per-execution `diagramContentHash` at ingest time (point-in-time link from `ExecutionDetail`/`ExecutionSummary`).
-`SensitiveKeysMerger` — pure function: merge(global, perApp) -> union with case-insensitive dedup, preserves first-seen casing. Returns null when both inputs null.
-`AppSettings`, `AppSettingsRepository` — per-app settings config and persistence
-`AppSettings`, `AppSettingsRepository` — per-app-per-env settings config and persistence. Record carries `(applicationId, environment, …)`; repository methods are `findByApplicationAndEnvironment`, `findByEnvironment`, `save`, `delete(appId, env)`. `AppSettings.defaults(appId, env)` produces a default instance scoped to an environment.
-`ThresholdConfig`, `ThresholdRepository` — alerting threshold config and persistence
-`AuditService` — audit logging facade
-`AuditRecord`, `AuditResult`, `AuditCategory`, `AuditRepository` — audit trail records and persistence
-`OutboundConnection` — record: id, tenantId, name, description, url, method, defaultHeaders, defaultBodyTmpl, tlsTrustMode, tlsCaPemPaths, hmacSecretCiphertext, auth, allowedEnvironmentIds, createdAt, createdBy (String user_id), updatedAt, updatedBy (String user_id). `isAllowedInEnvironment(envId)` returns true when allowed-envs list is empty OR contains the env.
-`OutboundAuth` — sealed interface + records: `None | Bearer(tokenCiphertext) | Basic(username, passwordCiphertext)`. Jackson `@JsonTypeInfo(use = DEDUCTION)` — wire shape has no discriminator, subtype inferred from fields.
-`OutboundAuthKind`, `OutboundMethod` — enums
-`OutboundConnectionRepository` — CRUD by (tenantId, id): save/findById/findByName/listByTenant/delete
-`OutboundConnectionService` — create/update/delete/get/list with uniqueness + narrow-envs + delete-if-referenced guards. `rulesReferencing(id)` stubbed in Plan 01 (returns `[]`); populated in Plan 02 against `AlertRuleRepository`.
-`ChunkAccumulator` — batches data for efficient flush
-`IngestionService` — diagram + metrics facade (`ingestDiagram`, `acceptMetrics`, `getMetricsBuffer`). Execution ingestion went through here via the legacy `RouteExecution` shape until `ChunkAccumulator` took over writes from the chunked pipeline — the `ingestExecution` path plus its `ExecutionStore.upsert` / `upsertProcessors` dependencies were removed.
-`ChunkAccumulator` — batches data for efficient flush; owns the execution write path (chunks → buffers → flush scheduler → `ClickHouseExecutionStore.insertExecutionBatch`).
-`WriteBuffer` — bounded ring buffer for async flush
-`BufferedLogEntry` — log entry wrapper with metadata
-`MergedExecution`, `TaggedExecution`, `TaggedDiagram` — tagged ingestion records
-`MergedExecution`, `TaggedDiagram` — tagged ingestion records. `TaggedDiagram` carries `(instanceId, applicationId, environment, graph)` — env is resolved from the agent registry in the controller and stamped on the ClickHouse `route_diagrams` row.
When deployed via the cameleer-saas platform, this server orchestrates customer app containers using Docker. Key components:
- **ConfigMerger** (`core/runtime/ConfigMerger.java`) — pure function: resolve(globalDefaults, envConfig, appConfig) -> ResolvedContainerConfig. Three-layer merge: global (application.yml) -> environment (defaultContainerConfig JSONB) -> app (containerConfig JSONB). Includes `runtimeType` (default `"auto"`) and `customArgs` (default `""`).
- **TraefikLabelBuilder** (`app/runtime/TraefikLabelBuilder.java`) — generates Traefik Docker labels for path-based (`/{envSlug}/{appSlug}/`) or subdomain-based (`{appSlug}-{envSlug}.{domain}`) routing. Supports strip-prefix and SSL offloading toggles. Also sets per-replica identity labels: `cameleer.replica` (index) and `cameleer.instance-id` (`{envSlug}-{appSlug}-{replicaIndex}`). Internal processing uses labels (not container name parsing) for extensibility.
- **TraefikLabelBuilder** (`app/runtime/TraefikLabelBuilder.java`) — generates Traefik Docker labels for path-based (`/{envSlug}/{appSlug}/`) or subdomain-based (`{appSlug}-{envSlug}.{domain}`) routing. Supports strip-prefix and SSL offloading toggles. Per-replica identity labels: `cameleer.replica` (index), `cameleer.generation` (8-char deployment UUID prefix — pin Prometheus/Grafana deploy boundaries with this), `cameleer.instance-id` (`{envSlug}-{appSlug}-{replicaIndex}-{generation}`). Traefik router/service keys deliberately omit the generation so load balancing spans old + new replicas during a blue/green overlap. When `ResolvedContainerConfig.externalRouting()` is `false` (UI: Resources → External Routing, default `true`), the builder emits ONLY the identity labels (`managed-by`, `cameleer.*`) and skips every `traefik.*` label — the container stays on `cameleer-traefik` and the per-env network (so sibling containers can still reach it via Docker DNS) but is invisible to Traefik. The `tls.certresolver` label is emitted only when `CAMELEER_SERVER_RUNTIME_CERTRESOLVER` is set to a non-blank resolver name (matching a resolver configured in the Traefik static config). When unset (dev installs backed by a static TLS store) only `tls=true` is emitted and Traefik serves the default cert from the TLS store.
- **PrometheusLabelBuilder** (`app/runtime/PrometheusLabelBuilder.java`) — generates Prometheus `docker_sd_configs` labels per resolved runtime type: Spring Boot `/actuator/prometheus:8081`, Quarkus/native `/q/metrics:9000`, plain Java `/metrics:9464`. Labels merged into container metadata alongside Traefik labels at deploy time.
- **DockerNetworkManager** (`app/runtime/DockerNetworkManager.java`) — manages two Docker network tiers:
-`cameleer-traefik` — shared network; Traefik, server, and all app containers attach here. Server joined via docker-compose with `cameleer-server` DNS alias.
-`cameleer-env-{slug}` — per-environment isolated network; containers in the same environment discover each other via Docker DNS. In SaaS mode, env networks are tenant-scoped: `cameleer-env-{tenantId}-{envSlug}` (overloaded `envNetworkName(tenantId, envSlug)` method) to prevent cross-tenant collisions when multiple tenants have identically-named environments.
- **DockerEventMonitor** (`app/runtime/DockerEventMonitor.java`) — persistent Docker event stream listener for containers with `managed-by=cameleer-server` label. Detects die/oom/start/stop events and updates deployment replica states. Periodic reconciliation (@Scheduled every 30s) inspects actual container state and corrects deployment status mismatches (fixes stale DEGRADED with all replicas healthy).
- **ContainerLogForwarder** (`app/runtime/ContainerLogForwarder.java`) — streams Docker container stdout/stderr to ClickHouse `logs` table with `source='container'`. Uses `docker logs --follow` per container, batches lines every 2s or 50 lines. Parses Docker timestamp prefix, infers log level via regex. `DeploymentExecutor` starts capture after each replica launches with the replica's `instanceId` (`{envSlug}-{appSlug}-{replicaIndex}`); `DockerEventMonitor` stops capture on die/oom. 60-second max capture timeout with 30s cleanup scheduler. Thread pool of 10 daemon threads. Container logs use the same `instanceId` as the agent (set via `CAMELEER_AGENT_INSTANCEID` env var) for unified log correlation at the instance level.
- **ContainerLogForwarder** (`app/runtime/ContainerLogForwarder.java`) — streams Docker container stdout/stderr to ClickHouse `logs` table with `source='container'`. Uses `docker logs --follow` per container, batches lines every 2s or 50 lines. Parses Docker timestamp prefix, infers log level via regex. `DeploymentExecutor` starts capture after each replica launches with the replica's `instanceId` (`{envSlug}-{appSlug}-{replicaIndex}-{generation}`); `DockerEventMonitor` stops capture on die/oom. 60-second max capture timeout with 30s cleanup scheduler. Thread pool of 10 daemon threads. Container logs use the same `instanceId` as the agent (set via `CAMELEER_AGENT_INSTANCEID` env var) for unified log correlation at the instance level. Instance-id changes per deployment — cross-deploy queries aggregate on `application + environment` (and optionally `replica_index`).
- **StartupLogPanel** (`ui/src/components/StartupLogPanel.tsx`) — collapsible log panel rendered below `DeploymentProgress`. Queries `/api/v1/logs?source=container&application={appSlug}&environment={envSlug}`. Auto-polls every 3s while deployment is STARTING; shows green "live" badge during polling, red "stopped" badge on FAILED. Uses `useStartupLogs` hook and `LogViewer` (design system).
## Container Hardening (issue #152)
`DockerRuntimeOrchestrator.startContainer` applies an unconditional hardening contract to every tenant container — Java 17 has no SecurityManager so the JVM is not a security boundary, and isolation must live below it. Defaults are fail-closed and have no opt-out:
-`cap_drop` = every `Capability.values()` (effectively ALL — docker-java's enum has no `ALL` constant). Outbound TCP still works (no caps needed); raw sockets, ptrace, mounts, and bind <1024 are denied.
-`security_opt`: `no-new-privileges:true`, `apparmor=docker-default`. Default seccomp profile is applied implicitly when `seccomp=` is absent.
-`read_only` rootfs = true.
-`pids_limit` = 512 (`PIDS_LIMIT` constant).
-`tmpfs` mount: `/tmp` with `rw,nosuid,size=256m`. **No `noexec`** — Netty/tcnative, Snappy, LZ4, Zstd dlopen native libs from `/tmp` via `mmap(PROT_EXEC)` which `noexec` blocks. Issue #153 will add per-app `writeableVolumes` for stateful tenants (Kafka Streams etc.).
**Sandboxed runtime auto-detect**: at construction the orchestrator calls `dockerClient.infoCmd().exec().getRuntimes()` and uses `runsc` (gVisor) when present. Override with `cameleer.server.runtime.dockerruntime` (e.g. `kata` to force Kata Containers, or any other registered runtime). Empty/blank = auto. The override always wins over auto-detect. The `DockerRuntimeOrchestrator(DockerClient, String)` constructor is the canonical entry point; the single-arg constructor exists only as a convenience for tests that don't need an override.
## DeploymentExecutor Details
Primary network for app containers is set via `CAMELEER_SERVER_RUNTIME_DOCKERNETWORK` env var (in SaaS mode: `cameleer-tenant-{slug}`); apps also connect to `cameleer-traefik` (routing) and `cameleer-env-{tenantId}-{envSlug}` (per-environment discovery) as additional networks. Resolves `runtimeType: auto` to concrete type from `AppVersion.detectedRuntimeType` at PRE_FLIGHT (fails deployment if unresolvable). Builds Docker entrypoint per runtime type (all JVM types use `-javaagent:/app/agent.jar -jar`, plain Java uses `-cp` with main class, native runs binary directly). Sets per-replica `CAMELEER_AGENT_INSTANCEID` env var to `{envSlug}-{appSlug}-{replicaIndex}` so container logs and agent logs share the same instance identity. Sets `CAMELEER_AGENT_*` env vars from `ResolvedContainerConfig` (routeControlEnabled, replayEnabled, health port). These are startup-only agent properties — changing them requires redeployment.
Primary network for app containers is set via `CAMELEER_SERVER_RUNTIME_DOCKERNETWORK` env var (in SaaS mode: `cameleer-tenant-{slug}`); apps also connect to `cameleer-traefik` (routing) and `cameleer-env-{tenantId}-{envSlug}` (per-environment discovery) as additional networks. Resolves `runtimeType: auto` to concrete type from `AppVersion.detectedRuntimeType` at PRE_FLIGHT (fails deployment if unresolvable). Builds Docker entrypoint per runtime type (all JVM types use `-javaagent:/app/agent.jar -jar`, plain Java uses `-cp` with main class, native runs binary directly). Sets per-replica `CAMELEER_AGENT_INSTANCEID` env var to `{envSlug}-{appSlug}-{replicaIndex}-{generation}` so container logs and agent logs share the same instance identity. Sets `CAMELEER_AGENT_*` env vars from `ResolvedContainerConfig` (routeControlEnabled, replayEnabled, health port). These are startup-only agent properties — changing them requires redeployment.
**Container naming** — `{tenantId}-{envSlug}-{appSlug}-{replicaIndex}-{generation}`, where `generation` is the first 8 characters of the deployment UUID. The generation suffix lets old + new replicas coexist during a blue/green swap (deterministic names without a generation used to 409). All lookups across the executor, `DockerEventMonitor`, and `ContainerLogForwarder` key on container **id**, not name — the name is operator-visibility only.
**Strategy dispatch** — `DeploymentStrategy.fromWire(config.deploymentStrategy())` branches the executor. Unknown values fall back to BLUE_GREEN so misconfiguration never throws at runtime.
- **Blue/green** (default): start all N new replicas → wait for ALL healthy → stop the previous deployment. Resource peak ≈ 2× replicas for the health-check window. Partial health aborts with status FAILED; the previous deployment is preserved untouched (user's safety net).
- **Rolling**: replace replicas one at a time — start new[i] → wait healthy → stop old[i] → next. Resource peak = replicas + 1. Mid-rollout health failure stops in-flight new containers and aborts; already-replaced old replicas are NOT restored (not reversible) but un-replaced old[i+1..N] keep serving traffic. User redeploys to recover.
Traffic routing is implicit: Traefik labels (`cameleer.app`, `cameleer.environment`) are generation-agnostic, so new replicas attract load balancing as soon as they come up healthy — no explicit swap step.
## Deployment Status Model
@@ -34,17 +55,13 @@ Primary network for app containers is set via `CAMELEER_SERVER_RUNTIME_DOCKERNET
| `STOPPED` | Intentionally stopped or initial state |
| `STARTING` | Deploy in progress |
| `RUNNING` | All replicas healthy and serving |
| `DEGRADED` | Some replicas healthy, some dead |
| `DEGRADED` | Post-deploy: a replica died after the deploy was marked RUNNING. Set by `DockerEventMonitor` reconciliation, never by `DeploymentExecutor` directly. |
| `STOPPING` | Graceful shutdown in progress |
| `FAILED` | Terminal failure (pre-flight, health check, or crash) |
| `FAILED` | Terminal failure (pre-flight, health check, or crash). Partial-healthy deploys now mark FAILED — DEGRADED is reserved for post-deploy drift. |
**Replica support**: deployments can specify a replica count. `DEGRADED` is used when at least one but not all replicas are healthy.
**Deploy stages** (`DeployStage`): PRE_FLIGHT -> PULL_IMAGE -> CREATE_NETWORK -> START_REPLICAS -> HEALTH_CHECK -> SWAP_TRAFFIC -> COMPLETE (or FAILED at any stage). Rolling reuses the same stage labels inside the per-replica loop; the UI progress bar shows the most recent stage.
**Deploy stages**(`DeployStage`): PRE_FLIGHT -> PULL_IMAGE -> CREATE_NETWORK -> START_REPLICAS -> HEALTH_CHECK -> SWAP_TRAFFIC -> COMPLETE (or FAILED at any stage).
**Blue/green strategy**: when re-deploying, new replicas are started and health-checked before old ones are stopped, minimising downtime.
**Deployment uniqueness**: `DeploymentService.createDeployment()` deletes any STOPPED/FAILED deployments for the same app+environment before creating a new one, preventing duplicate rows.
**Deployment retention**:`DeploymentService.createDeployment()` deletes FAILED deployments for the same app+environment before creating a new one, preventing failed-attempt buildup. STOPPED deployments are preserved as restorable checkpoints — the UI Checkpoints disclosure lists every deployment with a non-null `deployed_config_snapshot` (RUNNING, DEGRADED, STOPPED) minus the current one.
Server exposes `/api/v1/prometheus` (unauthenticated, Prometheus text format). Spring Boot Actuator provides JVM, GC, thread pool, and `http.server.requests` metrics automatically. Business metrics via `ServerMetrics` component:
Server exposes `/api/v1/prometheus` (unauthenticated, Prometheus text format). Spring Boot Actuator provides JVM, GC, thread pool, and `http.server.requests` metrics automatically. Business metrics via `ServerMetrics` component.
The same `MeterRegistry` is also snapshotted to ClickHouse every 60 s by `ServerMetricsSnapshotScheduler` (see "Server self-metrics persistence" at the bottom of this file) — so historical server-health data survives restarts without an external Prometheus.
## Gauges (auto-polled)
@@ -83,3 +85,23 @@ Mean processing time = `camel.route.policy.total_time / camel.route.policy.count
`ServerMetricsSnapshotScheduler` walks `MeterRegistry.getMeters()` every 60 s (configurable via `cameleer.server.self-metrics.interval-ms`) and writes one row per Micrometer `Measurement` to the ClickHouse `server_metrics` table. Full registry is captured — Spring Boot Actuator series (`jvm.*`, `process.*`, `http.server.requests`, `hikaricp.*`, `jdbc.*`, `tomcat.*`, `logback.events`, `system.*`) plus `cameleer.*` and `alerting_*`.
- No `environment` column — the server is env-agnostic.
-`tenant_id` threaded from `cameleer.server.tenant.id` (single-tenant per server).
-`server_instance_id` resolved once at boot by `ServerInstanceIdConfig` (property → HOSTNAME → localhost → UUID fallback). Rotates across restarts so counter resets are unambiguous.
- TTL: 90 days (vs 365 for `agent_metrics`). Write-only in v1 — no query endpoint or UI page. Inspect via ClickHouse admin: `/api/v1/admin/clickhouse/query` or direct SQL.
- Toggle off entirely with `cameleer.server.self-metrics.enabled=false` (uses `@ConditionalOnProperty`).
-Identity & Artifact section always visible; name editable pre-first-deploy, read-only after. JAR picker client-stages; new JAR + any form edits flip the primary button from `Save` to `Redeploy`. Environment fixed to the currently-selected env (no selector).
-Config sub-tabs: **Monitoring | Resources | Variables | Sensitive Keys | Deployment | ● Traces & Taps | ● Route Recording**. The four staged tabs feed dirty detection; the `●` live tabs apply in real-time (amber LiveBanner + default `?apply=live` on their writes) and never mark dirty.
- Primary action state machine: `Save` → `Uploading… N%` (during JAR upload; button shows percent with a tinted progress-fill overlay) → `Redeploy` → `Deploying…` during active deploy. Upload progress sourced from `useUploadJar` (XHR `upload.onprogress` → page-level `uploadPct` state). The button is disabled during `uploading` and `deploying`.
- Checkpoints render as a collapsible `CheckpointsTable` (default **collapsed**) **inside the Identity & Artifact `configGrid`** as an in-grid row (`Checkpoints | ▸ Expand (N)` / `▾ Collapse (N)`). `CheckpointsTable` returns a React.Fragment of grid-ready children so the label + trigger align with the other identity rows; when opened, a third grid child spans both columns via `grid-column: 1 / -1` so the 7-column table gets full width. Wired through `IdentitySection.checkpointsSlot` — `CheckpointDetailDrawer` stays in `IdentitySection.children` because it portals. Columns: Version · JAR (filename) · Deployed by · Deployed (relative `timeAgo` + user-locale sub-line via `new Date(iso).toLocaleString()`) · Strategy · Outcome · ›. Row click opens the drawer. Drawer tabs are ordered **Config | Logs** with `Config` as the default. Config panel has Snapshot / Diff vs current view modes. Replica filter in the Logs panel uses DS `Select`. Restore lives in the drawer footer (forces review). Visible row cap = `Environment.jarRetentionCount` (default 10 if 0/null); older rows accessible via "Show older (N)" expander. Currently-running deployment is excluded — represented separately by `StatusCard`. The empty-checkpoints case returns `null` (no row). The legacy `Checkpoints.tsx` row-list component is gone.
- Deployment tab: `StatusCard` + `DeploymentProgress` (during STARTING / FAILED) + flex-grow `StartupLogPanel` (no fixed maxHeight). Auto-activates when a deploy starts. The former `HistoryDisclosure` is retired — per-deployment config and logs live in the Checkpoints drawer. `StartupLogPanel` header mirrors the Runtime Application Log pattern: title + live/stopped badge + `N entries` + sort toggle (↑/↓, default **desc**) + refresh icon (`RefreshCw`). Sort drives the backend fetch via `useStartupLogs(…, sort)` so the 500-line limit returns the window closest to the user's interest; display order matches fetch order. Refresh scrolls to the latest edge (top for desc, bottom for asc). Sort + refresh buttons disable while a refetch is in flight. 3s polling while STARTING is unchanged.
- **Sensitive Keys** (`ui/src/pages/Admin/SensitiveKeysPage.tsx`) — global sensitive key masking config. Shows agent built-in defaults as outlined Badge reference, editable Tag pills for custom keys, amber-highlighted push-to-agents toggle. Keys add to (not replace) agent defaults. Per-app sensitive key additions managed via `ApplicationConfigController` API. Note: `AppConfigDetailPage.tsx` exists but is not routed in `router.tsx`.
- **Server Metrics** (`ui/src/pages/Admin/ServerMetricsAdminPage.tsx`) — dashboard over the `server_metrics` ClickHouse table. Visibility matches Database/ClickHouse pages: gated on `capabilities.infrastructureEndpoints` in `buildAdminTreeNodes`; backend is `@ConditionalOnProperty(infrastructureendpoints) + @PreAuthorize('hasRole(ADMIN)')`. Uses the generic `/api/v1/admin/server-metrics/{catalog,instances,query}` API via `ui/src/api/queries/admin/serverMetrics.ts` hooks (`useServerMetricsCatalog`, `useServerMetricsInstances`, `useServerMetricsSeries`), all three of which take a `ServerMetricsRange = { from: Date; to: Date }`. Time range is driven by the global TopBar picker via `useGlobalFilters()` — no page-local selector; bucket size auto-scales through `stepSecondsFor(windowSeconds)` (10 s up to 1 h buckets). Toolbar is just server-instance badges. Sections: Server health (agents/ingestion/auth), JVM (memory/CPU/GC/threads), HTTP & DB pools, Alerting (conditional on catalog), Deployments (conditional on catalog). Each panel is a `ThemedChart` with `Line`/`Area` children from the design system; multi-series responses are flattened into overlap rows by bucket timestamp. Alerting and Deployments rows are hidden when their metrics aren't in the catalog (zero-deploy / alerting-disabled installs).
## Key UI Files
@@ -25,11 +30,38 @@ The UI has 4 main tabs: **Exchanges**, **Dashboard**, **Runtime**, **Deployments
-`ui/src/components/ContentTabs.tsx` — main tab switcher
-`ui/src/components/EnvironmentSwitcherButton.tsx` + `EnvironmentSwitcherModal.tsx` — explicit env picker (button in TopBar; DS `Modal`-based list). Replaces the retired `EnvironmentSelector` (All-Envs dropdown). When `envRecords.length > 0` and the stored `selectedEnv` no longer matches any env, `LayoutShell` opens the modal in `forced` mode (non-dismissible). Switcher pulls env records from `useEnvironments()` (admin endpoint; readable by VIEWER+).
-`ui/src/components/env-colors.ts` + `ui/src/styles/env-colors.css` — 8-swatch preset palette for the per-environment color indicator. Tokens `--env-color-slate/red/amber/green/teal/blue/purple/pink` are defined for both light and dark themes. `envColorVar(name)` falls back to `slate` for unknown values. `LayoutShell` renders a 3px fixed top bar in the current env's color (z-index 900, below DS modals).
-`ui/src/components/StartupLogPanel.tsx` — deployment startup log viewer (container logs from ClickHouse, polls 3s while STARTING)
-`ui/src/api/queries/logs.ts` — `useStartupLogs` hook for container startup log polling, `useLogs`/`useApplicationLogs` for general log search
-`ui/src/api/queries/logs.ts` — `useStartupLogs` hook for container startup log polling, `useLogs`/`useApplicationLogs` for bounded log search (single page), `useInfiniteApplicationLogs` for streaming log views (cursor-paginated, server-side source/level filters)
-`ui/src/api/queries/agents.ts` — `useAgents` for agent list, `useInfiniteAgentEvents` for cursor-paginated timeline stream
-`ui/src/hooks/useInfiniteStream.ts` — tanstack `useInfiniteQuery` wrapper with top-gated auto-refetch, flattened `items[]`, and `refresh()` invalidator
-`ui/src/components/InfiniteScrollArea.tsx` — scrollable container with IntersectionObserver top/bottom sentinels. Streaming log/event views use this + `useInfiniteStream`. Bounded views (LogTab, StartupLogPanel) keep `useLogs`/`useStartupLogs`
-`ui/src/components/SideDrawer.tsx` — project-local right-slide drawer (DS has Modal but no Drawer). Portal-rendered, ESC + transparent-backdrop click closes, sticky header/footer, sizes md/lg/xl. Currently consumed only by `CheckpointDetailDrawer` — promote to `@cameleer/design-system` once a second consumer appears.
## Alerts
- **Sidebar section** (`buildAlertsTreeNodes` in `ui/src/components/sidebar-utils.ts`) — Inbox, Rules, Silences.
- **Routes** in `ui/src/router.tsx`: `/alerts` (redirect to inbox), `/alerts/inbox`, `/alerts/rules`, `/alerts/rules/new`, `/alerts/rules/:id`, `/alerts/silences`. No redirects for the retired `/alerts/all` and `/alerts/history` — stale URLs 404 per the clean-break policy.
- **Pages** under `ui/src/pages/Alerts/`:
-`InboxPage.tsx` — single filterable inbox. Filters: severity (multi), state (PENDING/FIRING/RESOLVED, default FIRING), Hide acked toggle (default on), Hide read toggle (default on). Row actions: Acknowledge, Mark read, Silence rule… (duration quick menu), Delete (OPERATOR+, soft-delete with undo toast wired to `useRestoreAlert`). Bulk toolbar (selection-driven): Acknowledge N · Mark N read · Silence rules · Delete N (ConfirmDialog; OPERATOR+).
-`SilenceRuleMenu.tsx` — DS `Dropdown`-based duration picker (1h / 8h / 24h / Custom…). Used by the row-level and bulk silence actions. "Custom…" navigates to `/alerts/silences?ruleId=<id>`.
-`RulesListPage.tsx` — CRUD + enable/disable toggle + env-promotion dropdown (pure UI prefill, no new endpoint).
-`RuleEditor/RuleEditorWizard.tsx` — 5-step wizard (Scope / Condition / Trigger / Notify / Review). `form-state.ts` is the single source of truth (`initialForm` / `toRequest` / `validateStep`). Seven condition-form subcomponents under `RuleEditor/condition-forms/` — including `AgentLifecycleForm.tsx` (multi-select event-type chips for the six-entry `AgentLifecycleEventType` allowlist + lookback-window input).
-`SilencesPage.tsx` — matcher-based create + end-early. Reads `?ruleId=` search param to prefill the Rule ID field (driven by InboxPage's "Silence rule… → Custom…" flow).
-`AlertRow.tsx` shared list row; `alerts-page.module.css` shared styling.
- **Components**:
-`NotificationBell.tsx` — polls `/alerts/unread-count` every 30 s (paused when tab hidden via TanStack Query `refetchIntervalInBackground: false`).
-`MustacheEditor/` — CodeMirror 6 editor with variable autocomplete + inline linter. Shared between rule title/message, webhook body/header overrides, and (future) Admin Outbound Connection editor (reduced-context mode for URL).
-`MustacheEditor/alert-variables.ts` — variable registry aligned with `NotificationContextBuilder.java`. Add new leaves here whenever the backend context grows.
- **API queries** under `ui/src/api/queries/`: `alerts.ts`, `alertRules.ts`, `alertSilences.ts`, `alertNotifications.ts`, `alertMeta.ts`. All env-scoped via `useSelectedEnv` from `alertMeta`.
- **CMD-K**: `buildAlertSearchData` in `LayoutShell.tsx` registers `alert` and `alertRule` result categories. Badges convey severity + state. Palette navigates directly to the deep-link path — no sidebar-reveal state for alerts.
- **Top-nav**: `<NotificationBell />` is the first child of `<TopBar>`, sitting alongside `SearchTrigger` + status `ButtonGroup` + `TimeRangeDropdown` + `AutoRefreshToggle`.
Branch: `main`, starting HEAD `90460705` (chore: refresh GitNexus index stats).
## Summary
- **Starting state**: 65 IT failures (46 F + 19 E) out of 555 tests on a clean build. Side-note: `target/classes` incremental-build staleness from the `90083f88` V1..V18 → V1 schema collapse makes the number look worse (every context load dies on `Flyway V2__claim_mapping.sql failed`). A fresh `mvn clean verify` gives the real 65.
- **Final state**: **12 failures across 3 test classes** (`AgentSseControllerIT`, `SseSigningIT`, `ClickHouseStatsStoreIT`). **53 failures closed across 14 test classes.**
- **11 commits landed on local `main`** (not pushed).
- No new env vars, endpoints, tables, or columns added. `V1__init.sql` untouched. No tests rewritten to pass-by-weakening — every assertion change is accompanied by a comment explaining the contract it now captures.
| `b55221e9` | SensitiveKeysAdminControllerIT | assert pushResult shape, not exact 0 (shared registry across ITs) |
## The single biggest insight
**`ExecutionController` (legacy PG path) is dead code.** It's `@ConditionalOnMissingBean(ChunkAccumulator.class)` and `ChunkAccumulator` is registered **unconditionally** in `StorageBeanConfig.java:92`, so `ExecutionController` never binds. Even if it did, `IngestionService.upsert` → `ClickHouseExecutionStore.upsert` throws `UnsupportedOperationException("ClickHouse writes use the chunked pipeline")` — the only `ExecutionStore` impl in `src/main/java` is ClickHouse, the Postgres variant lives in a planning doc only.
Practical consequences for every IT that was exercising `/api/v1/data/executions`:
1.`ChunkIngestionController` owns the URL and expects an `ExecutionChunk` envelope (`exchangeId`, `applicationId`, `instanceId`, `routeId`, `status`, `startTime`, `endTime`, `durationMs`, `chunkSeq`, `final`, `processors: FlatProcessorRecord[]`) — the legacy `RouteExecution` shape was being silently degraded to an empty/degenerate chunk.
2. The test payload changes are accompanied by assertion changes that now go through REST endpoints instead of raw SQL against the (ClickHouse-resident) `executions` / `processor_executions` / `route_diagrams` / `agent_metrics` tables.
3.**Recommendation for cleanup**: remove `ExecutionController` + the `upsert` path in `IngestionService` + the stubbed `ClickHouseExecutionStore.upsert` throwers. Separate PR. Happy to file.
## Cluster breakdown
**Cluster A — missing `environmentId` in register bodies (DONE)**
Root cause: `POST /api/v1/agents/register` now 400s without `environmentId`. Test payloads minted before this requirement. Fixed across all agent-registering ITs plus side-cleanups (flaky iat-coupled assertion in JwtRefreshIT, wrong RBAC target in can-access tests, absolute vs relative sseEndpoint).
**Cluster B — ingestion payload drift (DONE per user direction)**
All controller + storage ITs that posted `RouteExecution` JSON now post `ExecutionChunk` envelopes. All CH-side assertions now go through REST endpoints (`/api/v1/environments/{env}/executions` search + `/api/v1/executions/{id}` detail + `/agents/{id}/metrics` + `/apps/{app}/routes/{route}/diagram`). DiagramRenderControllerIT's SVG tests still need a content hash → reads it off the execution-detail REST response rather than querying `route_diagrams`.
**Cluster C — flat URL drift (DONE)**
`/api/v1/agents` → `/api/v1/environments/{envSlug}/agents`. Two test classes touched.
**Cluster D — heartbeat auto-heal contract (DONE)**
`heartbeatUnknownAgent_returns404` renamed and asserts the 200 auto-heal path that `fb54f9cb` made the contract.
**Cluster E — individual drifts (DONE except three parked)**
| Test class | Status |
|---|---|
| FlywayMigrationIT | DONE (decouple from shared PG state) |
| ClickHouseExecutionReadIT | DONE (iteration vs loopIndex mapping) |
## PARKED — what you'll want to look at next
### 1. ClickHouseStatsStoreIT (8 failures) — timezone bug in production code
`ClickHouseStatsStore.buildStatsSql` uses `lit(Instant)` which formats as `'yyyy-MM-dd HH:mm:ss'` in UTC but with no timezone marker. ClickHouse parses that literal in the session timezone when comparing against the `DateTime`-typed `bucket` column in `stats_1m_*`. On a non-UTC CH host (e.g. CEST docker on a CEST laptop), the filter endpoint is off by the tz offset in hours and misses every row the MV bucketed.
I confirmed this by instrumenting the test: `toDateTime(bucket)` returned `12:00:00` for a row inserted with `start_time=10:00:00Z` (i.e. the stored UTC Unix timestamp but displayed in CEST), and the filter literal `'2026-03-31 10:05:00'` was being parsed as CEST → UTC 08:05 → excluded all rows.
**I didn't fix this** because the repair is in `src/main/java`, not the test. Two reasonable options:
- **Test-side**: pin the container TZ via `.withEnv("TZ", "UTC")` + include `use_time_zone=UTC` in the JDBC URL. I tried both; neither was sufficient on their own — the CH server reads `timezone` from its own config, not `$TZ`. Getting all three layers (container env, CH server config, JDBC driver) aligned needs dedicated effort.
- **Production-side (preferred)**: change `lit(Instant)` to `toDateTime('...', 'UTC')` or use the 3-arg `DateTime(3, 'UTC')` column type for `bucket`. That's a store change; would be caught by a matching unit test.
I did add the explicit `'default'` env to the seed `INSERT`s per your directive, but reverted it locally because the timezone bug swallowed the fix. The raw unchanged test is what's committed.
All failing assertions are `awaitConnection(5000)` timeouts or `ConditionTimeoutException` on SSE stream observation. Not related to any spec drift I could identify — the SSE server is up (other tests in the same classes connect fine), and auth/JWT is accepted. Looks like a real race on either the SseConnectionManager registration or on the HTTP client's first-read flush. Needs a dedicated debug session with a minimal reproducer; not something I wanted to hack around with sleeps.
Specific tests:
-`AgentSseControllerIT.sseConnect_unknownAgent_returns404` — 5s `CompletableFuture.get` timeout on an HTTP GET that should return 404 synchronously. Suggests the client is waiting on body data that never arrives (SSE stream opens even on 404?).
-`AgentSseControllerIT.pingKeepalive_receivedViaSseStream` — waits for an event line in the stream snapshot, never sees it.
-`SseSigningIT.deepTraceEvent_containsValidSignature` — same pattern.
The sibling tests (`SseSigningIT.configUpdateEvent_containsValidEd25519Signature`) pass in isolation, which strongly suggests order-dependent flakiness rather than a protocol break.
Reports land in `cameleer-server-app/target/failsafe-reports/`. Expect **12 failures** in the three classes above. Everything else is green.
## Side notes worth flagging
- **Property-key inconsistency in the main code** — surfaced via BackpressureIT. `IngestionConfig` is bound under `cameleer.server.ingestion.*`, but `MetricsFlushScheduler.@Scheduled` reads `ingestion.flush-interval-ms` (no prefix, hyphenated). In production this means the flush-interval in `application.yml` isn't actually being honoured by the metrics flush — it stays at the 1s fallback. Separate cleanup.
- **Shared Testcontainers PG across IT classes** — several of the "cross-test state" fixes (FlywayMigrationIT, ConfigEnvIsolationIT, SensitiveKeysAdminControllerIT) are symptoms of one underlying issue: `AbstractPostgresIT` uses a singleton PG container, and nothing cleans between test classes. Could do with a global `@Sql("/test-reset.sql")` on `@BeforeAll`, but out of scope here.
- **Agent registry shared across ITs** — same class of issue. Doesn't bite until a test explicitly inspects registry membership (SensitiveKeys `pushResult.total`).
All three parked clusters now green. 560/560 tests passing.
- **ClickHouseStatsStoreIT (8 failures)** — fixed in `a9a6b465`. Two-layer TZ fix: JVM default TZ pinned to UTC in `CameleerServerApplication.main()` (the ClickHouse JDBC 0.9.7 driver formats `java.sql.Timestamp` via `Timestamp.toString()`, which uses JVM default TZ — a CEST JVM shipping to a UTC CH server stored off-by-offset Unix timestamps), plus column-level `bucket DateTime('UTC')` on all `stats_1m_*` tables with explicit `toDateTime(..., 'UTC')` casts in MV projections and `ClickHouseStatsStore.lit(Instant)` as defence in depth.
- **MetricsFlushScheduler property-key drift** — fixed in `a6944911`. Scheduler now reads `${cameleer.server.ingestion.flush-interval-ms:1000}` (the SpEL-via-`@ingestionConfig` approach doesn't work because `@EnableConfigurationProperties` uses a compound bean name). BackpressureIT workaround property removed.
- **SSE flakiness (4 failures, `AgentSseControllerIT` + `SseSigningIT`)** — fixed in `41df042e`. Triage's "order-dependent flakiness" theory was wrong — all four reproduced in isolation. Three root causes: (a) `AgentSseController.events` auto-heal was over-permissive (spoofing vector), fixed with JWT-subject-equals-path-id check; (b) `SseConnectionManager.pingAll` read an unprefixed property key (`agent-registry.ping-interval-ms`), same family of bug as (a6944911); (c) SSE response headers didn't flush until the first `emitter.send()`, so `awaitConnection(5s)` assertions timed out under the 15s ping cadence — fixed by sending an initial `: connected` comment on `connect()`. Full diagnosis in `.planning/sse-flakiness-diagnosis.md`.
Plus the two prod-code cleanups from the ExecutionController-removal follow-ons:
- **Dead `SearchIndexer` subsystem** — removed in `98cbf8f3`. `ExecutionUpdatedEvent` had no publisher after `0f635576`, so the whole indexer + stats + `/admin/clickhouse/pipeline` endpoint + UI pipeline card carried zero signal.
- **Unused `TaggedExecution` record** — removed in `06c6f53b`.
Not order-dependent flakiness (triage report was wrong). Three distinct root causes, one production bug and one test-infrastructure issue, all reproducible when running the classes in isolation.
**Bug:** auto-heal registers *any* path id when any valid JWT is present, regardless of whether the JWT subject matches the path id. A holder of agent X's JWT can open SSE for any path-id Y, silently spoofing Y.
**Surface symptom:**`sseConnect_unknownAgent_returns404` sends a JWT for `test-agent-sse-it` and requests SSE for `unknown-sse-agent`. Auto-heal kicks in, returns 200 with an infinite empty stream. Test's `statusFuture.get(5s)` — which uses `BodyHandlers.ofString()` and waits for the full body — times out instead of getting a synchronous 404.
**Fix:** only auto-heal when `jwtResult.subject().equals(id)`.
### 2. `SseConnectionManager.pingAll` reads an unprefixed property key (production bug)
**Bug:**`AgentRegistryConfig` is `@ConfigurationProperties(prefix = "cameleer.server.agentregistry")`. The scheduler reads an unprefixed `agent-registry.*` key that the YAML never defines — so the default 15s always applies, regardless of config. Same family of bug as the `MetricsFlushScheduler` fix in commit `a6944911`.
Spring's `SseEmitter` holds the response open but doesn't flush headers to the client until the first `emitter.send()`. Until then, clients using `HttpResponse.BodyHandlers.ofInputStream()` block on the first byte.
**Surface symptom:**
-`lastEventIdHeader_connectionSucceeds` — asserts `awaitConnection(5000)` is `true`. The latch counts down in `.thenAccept(response -> ...)`, which in practice only fires once body bytes start flowing (JDK 21 behaviour with SSE streams). Default ping cadence is 15s → 5s assertion times out.
-`pingKeepalive_receivedViaSseStream` — waits 5s for a `:ping` line. The scheduler runs every 15s (both by default, and because of bug #2, unconditionally).
-`SseSigningIT.deepTraceEvent_containsValidSignature` — same family: `awaitConnection(5000).isTrue()`.
**Fix:** send an initial `: connected` comment as part of `connect()`. Spring flushes on the first `.send()`, so an immediate comment forces the response headers + first byte to hit the wire, which triggers the client's `thenAccept` callback. Also solves the ping-test: the initial comment is observed as a keepalive line within the test's polling window.
## Hypothesis ladder (ruled out)
- **Order-dependent singleton leak** — ruled out: every failure reproduces when the class is run solo.
- **Tomcat async thread pool exhaustion** — ruled out: `SseEmitter(Long.MAX_VALUE)` does hold threads, but the 7-test class doesn't reach Tomcat's defaults.
- **SseConnectionManager emitter-map contamination** — ruled out: each test uses a unique agent id (UUID-suffixed), and the `@Component` is the same instance across tests but the emitter map is keyed by agent id, no collisions.
This project is indexed by GitNexus as **cameleer-server** (6306 symbols, 15892 relationships, 300 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely.
This project is indexed by GitNexus as **cameleer-server** (9731 symbols, 24987 relationships, 300 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely.
> If any GitNexus tool warns the index is stale, run `npx gitnexus analyze` in terminal first.
@@ -14,16 +14,29 @@ Cameleer Server — observability server that receives, stores, and serves Camel
## Modules
-`cameleer-license-api` — pure license contract types (`LicenseInfo`, `LicenseValidator`, `LicenseState`, `LicenseStateMachine`, `LicenseLimits`, `DefaultTierLimits`) under package `com.cameleer.license`. No Spring or server-runtime deps; consumed by `cameleer-server-core` (validation/runtime gate) and `cameleer-license-minter` (vendor signing) — and transitively by `cameleer-saas` via the minter — without inheriting server internals.
-`cameleer-server-core` — domain logic, storage interfaces, services (no Spring dependencies)
-`cameleer-server-app` — Spring Boot web app, REST controllers, SSE, persistence, Docker orchestration
-`cameleer-license-minter` — vendor-only Ed25519 license signing library + CLI. Depends only on `cameleer-license-api` so consumers don't pull in `cameleer-server-core`.
## Build Commands
```bash
mvn clean compile # Compile all modules
mvn clean verify # Full build with tests
mvn clean verify -DskipITs # Fast: unit tests only (no Testcontainers)
```
### Faster local builds
- **Surefire reuses forks** (`cameleer-server-app/pom.xml`): unit tests run with `forkCount=1C` + `reuseForks=true` — one JVM per CPU core, reused across classes. Test classes that mutate static state must clean up after themselves.
- **Testcontainers reuse** — opt-in per developer. Add to `~/.testcontainers.properties`:
```
testcontainers.reuse.enable=true
```
Then `AbstractPostgresIT` containers persist across `mvn verify` runs (saves ~20s per run). Stop them manually when you need a clean DB: `docker rm -f $(docker ps -aq --filter label=org.testcontainers.reuse=true)`.
- **UI build** dropped redundant `tsc --noEmit` from `npm run build` (Vite/esbuild type-checks during bundling). Run `npm run typecheck` explicitly when you want a full type-check pass.
- Depends on `com.cameleer:cameleer-common` from Gitea Maven registry
- Jackson `JavaTimeModule` for `Instant` deserialization
- Communication: receives HTTP POST data from agents (executions, diagrams, metrics, logs), serves SSE event streams for config push/commands (config-update, deep-trace, replay, route-control)
- Environment filtering: all data queries filter by the selected environment. All commands target only agents in the selected environment. Backend endpoints accept optional `environment` query parameter; null = all environments (backward compatible).
- Maintains agent instance registry (in-memory) with states: LIVE -> STALE -> DEAD. Auto-heals from JWT `env` claim + heartbeat body on heartbeat/SSE after server restart (priority: heartbeat `environmentId` > JWT `env` claim > `"default"`). Capabilities and route states updated on every heartbeat (protocol v2). Route catalog falls back to ClickHouse stats for route discovery when registry has incomplete data.
- URL taxonomy: user-facing data, config, and query endpoints live under `/api/v1/environments/{envSlug}/...`. Env is a path segment, resolved via the `@EnvPath` argument resolver (404 on unknown slug). Flat endpoints are only for: agent self-service (JWT-authoritative), cross-env admin (RBAC, OIDC, audit, license, thresholds, env CRUD), cross-env discovery (`/catalog`), content-addressed lookups (`/diagrams/{contentHash}/render`, `/executions/{id}`), and auth. See `.claude/rules/app-classes.md` for the full allow-list.
- Slug immutability: environment and app slugs are immutable after creation (both appear in URLs, Docker network names, container names, and ClickHouse partition keys). Slug regex `^[a-z0-9][a-z0-9-]{0,63}$` is enforced on POST; update endpoints silently drop any slug field in the request body via Jackson's default unknown-property handling.
- App uniqueness: `(environment_id, app_slug)` is the natural key. The same app slug can legitimately exist in multiple environments; `AppService.getByEnvironmentAndSlug(envId, slug)` is the canonical lookup for controllers. Bare `getBySlug(slug)` remains for internal use but is ambiguous across envs.
- Environment filtering: all data queries filter by the selected environment. All commands target only agents in the selected environment. Env is required on every env-scoped endpoint (path param); the legacy `?environment=` query form is retired.
- Maintains agent instance registry (in-memory) with states: LIVE -> STALE -> DEAD. Auto-heals from JWT `env` claim + heartbeat body on heartbeat/SSE after server restart (priority: heartbeat `environmentId` > JWT `env` claim; no silent default — missing env on heartbeat auto-heal returns 400). Registration (`POST /api/v1/agents/register`) requires `environmentId` in the request body; missing or blank returns 400. Capabilities and route states updated on every heartbeat (protocol v2). Route catalog merges three sources: in-memory agent registry, persistent `route_catalog` table (ClickHouse), and `stats_1m_route` execution stats. The persistent catalog tracks `first_seen`/`last_seen` per route per environment, updated on every registration and heartbeat. Routes appear in the sidebar when their lifecycle overlaps the selected time window (`first_seen <= to AND last_seen >= from`), so historical routes remain visible even after being dropped from newer app versions.
- Multi-tenancy: each server instance serves one tenant (configured via `CAMELEER_SERVER_TENANT_ID`, default: `"default"`). Environments (dev/staging/prod) are first-class. PostgreSQL isolated via schema-per-tenant (`?currentSchema=tenant_{id}`) and `ApplicationName=tenant_{id}` on the JDBC URL. ClickHouse shared DB with `tenant_id` + `environment` columns, partitioned by `(tenant_id, toYYYYMM(timestamp))`.
- Storage: PostgreSQL for RBAC, config, and audit; ClickHouse for all observability data (executions, search, logs, metrics, stats, diagrams). ClickHouse schema migrations in `clickhouse/*.sql`, run idempotently on startup by `ClickHouseSchemaInitializer`. Use `IF NOT EXISTS` for CREATE and ADD PROJECTION.
- Log exchange correlation: `ClickHouseLogStore` extracts `exchange_id` from log entry MDC, preferring `cameleer.exchangeId` over `camel.exchangeId` (fallback for older agents). For `ON_COMPLETION` exchange copies, the agent sets `cameleer.exchangeId` to the parent's exchange ID via `CORRELATION_ID`.
- Log processor correlation: The agent sets `cameleer.processorId` in MDC, identifying which processor node emitted a log line.
- Logging: ClickHouse JDBC set to INFO (`com.clickhouse`), HTTP client to WARN (`org.apache.hc.client5`) in application.yml
- Security: JWT auth with RBAC (AGENT/VIEWER/OPERATOR/ADMIN roles), Ed25519 config signing (key derived deterministically from JWT secret via HMAC-SHA256), bootstrap token for registration. CORS: `CAMELEER_SERVER_SECURITY_CORSALLOWEDORIGINS` (comma-separated) overrides `CAMELEER_SERVER_SECURITY_UIORIGIN` for multi-origin setups. Infrastructure access: `CAMELEER_SERVER_SECURITY_INFRASTRUCTUREENDPOINTS=false` disables Database and ClickHouse admin endpoints. Last-ADMIN guard: system prevents removal of the last ADMIN role (409 Conflict). Password policy: min 12 chars, 3-of-4 character classes, no username match. Brute-force protection: 5 failed attempts -> 15 min lockout. Token revocation: `token_revoked_before` column on users, checked in `JwtAuthenticationFilter`, set on password change.
- Login routing: `GET /api/v1/auth/capabilities` (unauthenticated) tells the SPA whether OIDC is the primary entry point. When OIDC is configured, the SSO button is the primary CTA and the local form is hidden behind `?local` (admin-recovery escape hatch). Per RFC 9700 §4.4 we do **not** use `prompt=none` for primary login — that returns `login_required` for first-time users and traps them on a local form.
- OIDC: Optional external identity provider support (token exchange pattern). Configured via admin API/UI, stored in database (`server_config` table). Resource server mode: accepts external access tokens (Logto M2M) via JWKS validation when `CAMELEER_SERVER_SECURITY_OIDCISSUERURI` is set. Scope-based role mapping via `SystemRole.normalizeScope()`. System roles synced on every OIDC login via `applyClaimMappings()` in `OidcAuthController` (calls `clearManagedAssignments` + `assignManagedRole` on `RbacService`) — always overwrites managed role assignments; uses managed assignment origin to avoid touching group-inherited or directly-assigned roles. Supports ES384, ES256, RS256.
- OIDC role extraction: `OidcTokenExchanger` reads roles from the **access_token** first (JWT with `at+jwt` type), then falls back to id_token. `OidcConfig` includes `audience` (RFC 8707 resource indicator) and `additionalScopes`. All provider-specific configuration is external — no provider-specific code in the server.
- Sensitive keys: Global enforced baseline for masking sensitive data in agent payloads. Merge rule: `final = global UNION per-app` (case-insensitive dedup, per-app can only add, never remove global keys).
- User persistence: PostgreSQL `users` table, admin CRUD at `/api/v1/admin/users`
- User persistence: PostgreSQL `users` table, admin CRUD at `/api/v1/admin/users`. `users.user_id` is the **bare** identifier — local users as `<username>`, OIDC users as `oidc:<sub>`. JWT `sub` carries the `user:` namespace prefix so `JwtAuthenticationFilter` can tell user tokens from agent tokens; write paths (`UiAuthController`, `OidcAuthController`, `UserAdminController`) all upsert unprefixed, and env-scoped read-path controllers strip the `user:` prefix before using the value as an FK to `users.user_id` / `user_roles.user_id`. Alerting / outbound FKs (`alert_rules.created_by`, `outbound_connections.created_by`, …) therefore all reference the bare form.
- V8 — Deployment active config (resolved_config JSONB on deployments)
- V9 — Password hardening (failed_login_attempts, locked_until, token_revoked_before on users)
- V10 — Runtime type detection (detected_runtime_type, detected_main_class on app_versions)
- V1 — Consolidated baseline schema. All prior V1–V18 evolution was collapsed before first prod deploy. Contains: RBAC (users, roles, groups, user_roles, user_groups, group_roles, claim_mapping_rules), runtime management (environments, apps, app_versions, deployments), env-scoped application config (application_config PK `(application, environment)`, app_settings PK `(application_id, environment)`), audit_log, outbound_connections, server_config, and the full alerting subsystem (alert_rules, alert_rule_targets, alert_instances, alert_silences, alert_notifications). Seeds the 4 system roles (AGENT/VIEWER/OPERATOR/ADMIN), the `Admins` group with ADMIN role, and a default environment. Invariants covered by `SchemaBootstrapIT`.
ClickHouse: `cameleer-server-app/src/main/resources/clickhouse/init.sql` (run idempotently on startup)
## Regenerating OpenAPI schema (SPA types)
After any change to REST controller paths, request/response DTOs, or `@PathVariable`/`@RequestParam`/`@RequestBody` signatures, regenerate the TypeScript types the SPA consumes. Required for every controller-level change.
```bash
# Backend must be running on :8081
cd ui && npm run generate-api:live # fetches fresh openapi.json AND regenerates schema.d.ts
# OR, if openapi.json was updated by other means:
cd ui && npm run generate-api # regenerates schema.d.ts from existing openapi.json
```
After regeneration, `ui/src/api/schema.d.ts` and `ui/src/api/openapi.json` will update. The TypeScript compiler then surfaces every SPA call site that needs updating — fix all compile errors before testing in the browser. Commit the regenerated files with the controller change.
## Maintaining .claude/rules/
When adding, removing, or renaming classes, controllers, endpoints, UI components, or metrics, update the corresponding `.claude/rules/` file as part of the same change. The rule files are the class/API map that future sessions rely on — stale rules cause wrong assumptions. Treat rule file updates like updating an import: part of the change, not a separate task.
@@ -74,3 +95,105 @@ When adding, removing, or renaming classes, controllers, endpoints, UI component
## Disabled Skills
- Do NOT use any `gsd:*` skills in this project. This includes all `/gsd:` prefixed commands.
<!-- gitnexus:start -->
# GitNexus — Code Intelligence
This project is indexed by GitNexus as **cameleer-server** (10530 symbols, 27383 relationships, 300 execution flows). Use the GitNexus MCP tools to understand code, assess impact, and navigate safely.
> If any GitNexus tool warns the index is stale, run `npx gitnexus analyze` in terminal first.
## Always Do
- **MUST run impact analysis before editing any symbol.** Before modifying a function, class, or method, run `gitnexus_impact({target: "symbolName", direction: "upstream"})` and report the blast radius (direct callers, affected processes, risk level) to the user.
- **MUST run `gitnexus_detect_changes()` before committing** to verify your changes only affect expected symbols and execution flows.
- **MUST warn the user** if impact analysis returns HIGH or CRITICAL risk before proceeding with edits.
- When exploring unfamiliar code, use `gitnexus_query({query: "concept"})` to find execution flows instead of grepping. It returns process-grouped results ranked by relevance.
- When you need full context on a specific symbol — callers, callees, which execution flows it participates in — use `gitnexus_context({name: "symbolName"})`.
## When Debugging
1. `gitnexus_query({query: "<error or symptom>"})` — find execution flows related to the issue
2. `gitnexus_context({name: "<suspect function>"})` — see all callers, callees, and process participation
3. `READ gitnexus://repo/cameleer-server/process/{processName}` — trace the full execution flow step by step
4. For regressions: `gitnexus_detect_changes({scope: "compare", base_ref: "main"})` — see what your branch changed
## When Refactoring
- **Renaming**: MUST use `gitnexus_rename({symbol_name: "old", new_name: "new", dry_run: true})` first. Review the preview — graph edits are safe, text_search edits need manual review. Then run with `dry_run: false`.
- **Extracting/Splitting**: MUST run `gitnexus_context({name: "target"})` to see all incoming/outgoing refs, then `gitnexus_impact({target: "target", direction: "upstream"})` to find all external callers before moving code.
- After any refactor: run `gitnexus_detect_changes({scope: "all"})` to verify only expected files changed.
## Never Do
- NEVER edit a function, class, or method without first running `gitnexus_impact` on it.
- NEVER ignore HIGH or CRITICAL risk warnings from impact analysis.
- NEVER rename symbols with find-and-replace — use `gitnexus_rename` which understands the call graph.
- NEVER commit changes without running `gitnexus_detect_changes()` to check affected scope.
Before completing any code modification task, verify:
1. `gitnexus_impact` was run for all modified symbols
2. No HIGH/CRITICAL risk warnings were ignored
3. `gitnexus_detect_changes()` confirms changes match expected scope
4. All d=1 (WILL BREAK) dependents were updated
## Keeping the Index Fresh
After committing code changes, the GitNexus index becomes stale. Re-run analyze to update it:
```bash
npx gitnexus analyze
```
If the index previously included embeddings, preserve them by adding `--embeddings`:
```bash
npx gitnexus analyze --embeddings
```
To check whether embeddings exist, inspect `.gitnexus/meta.json` — the `stats.embeddings` field shows the count (0 means no embeddings). **Running analyze without `--embeddings` will delete any previously generated embeddings.**
> Claude Code users: A PostToolUse hook handles this automatically after `git commit` and `git merge`.
## CLI
| Task | Read this skill file |
|------|---------------------|
| Understand architecture / "How does X work?" | `.claude/skills/gitnexus/gitnexus-exploring/SKILL.md` |
| Blast radius / "What breaks if I change X?" | `.claude/skills/gitnexus/gitnexus-impact-analysis/SKILL.md` |
| Trace bugs / "Why is X failing?" | `.claude/skills/gitnexus/gitnexus-debugging/SKILL.md` |
@@ -19,38 +19,99 @@ mvn clean compile # compile only
mvn clean verify # compile + run all tests (needs Docker for integration tests)
```
## Infrastructure Setup
## Start a brand-new local environment (Docker)
Start PostgreSQL:
The repo ships a `docker-compose.yml` with the full stack: PostgreSQL, ClickHouse, the Spring Boot server, and the nginx-served SPA. All dev defaults are baked into the compose file — no `.env` file or extra config needed for a first run.
```bash
# 1. Clean slate (safe if this is already a first run — noop when no volumes exist)
docker compose down -v
# 2. Build + start everything. First run rebuilds both images (~2–4 min).
docker compose up -d --build
# 3. Watch the server come up (health check goes green in ~60–90s after Flyway + ClickHouse init)
docker compose logs -f cameleer-server
# ready when you see "Started CameleerServerApplication in ...".
| `CAMELEER_SERVER_RUNTIME_ENABLED` | `false` (Docker-in-Docker app orchestration off for the local stack) |
Override any of these by editing `docker-compose.yml` or passing `-e KEY=value` to `docker compose run`.
### Common lifecycle commands
```bash
# Stop everything but keep volumes (quick restart later)
docker compose stop
# Start again after a stop
docker compose start
# Apply changes to the server code / UI — rebuild just what changed
docker compose up -d --build cameleer-server
docker compose up -d --build cameleer-ui
# Wipe the environment completely (drops PG + ClickHouse volumes — all data gone)
docker compose down -v
# Fresh Flyway run by dropping just the PG volume (keeps ClickHouse data)
docker compose down
docker volume rm cameleer-server_cameleer-pgdata
docker compose up -d
```
This starts PostgreSQL 16. The database schema is applied automatically via Flyway migrations on server startup. ClickHouse tables are created by the schema initializer on startup.
### Infra-only mode (backend via `mvn` / UI via Vite)
If you want to iterate on backend/UI code without rebuilding the server image on every change, start just the databases and run the server + UI locally:
```bash
# 1. Only infra containers
docker compose up -d cameleer-postgres cameleer-clickhouse
# 2. Build and run the server jar against those containers
# 3. In another terminal — Vite dev server on :5173 (proxies /api → :8081)
cd ui && npm install && npm run dev
```
> **Note:** The Docker image no longer includes default database credentials. When running via `docker run`, pass `-e SPRING_DATASOURCE_URL=...` etc. The docker-compose setup provides these automatically.
Database schema is applied automatically: PostgreSQL via Flyway migrations on server startup, ClickHouse tables via `ClickHouseSchemaInitializer`. No manual DDL needed.
The server starts on **port 8081**. The `CAMELEER_SERVER_SECURITY_BOOTSTRAPTOKEN` environment variable is **required** — the server fails fast on startup if it is not set.
For token rotation without downtime, set `CAMELEER_SERVER_SECURITY_BOOTSTRAPTOKENPREVIOUS` to the old token while rolling out the new one. The server accepts both during the overlap window.
`CAMELEER_SERVER_SECURITY_BOOTSTRAPTOKEN` is **required** for agent registration — the server fails fast on startup if it's not set. For token rotation without downtime, set `CAMELEER_SERVER_SECURITY_BOOTSTRAPTOKENPREVIOUS` to the old token while rolling out the new one — the server accepts both during the overlap window.
## API Endpoints
@@ -433,11 +494,13 @@ Key settings in `cameleer-server-app/src/main/resources/application.yml`. All cu
| `cameleer.server.runtime.dockerruntime` | *(empty = auto)* | `CAMELEER_SERVER_RUNTIME_DOCKERRUNTIME` | Container runtime override. Empty auto-detects gVisor (`runsc`) when registered with the daemon and falls back to the daemon default. Set to e.g. `kata` to force a specific runtime, or `runc` to force the default even if `runsc` is installed. |
<description>Pure license contract types — LicenseInfo, LicenseValidator, LicenseState, LicenseStateMachine, LicenseLimits, DefaultTierLimits. Shared by server-core (validation/runtime gate) and cameleer-license-minter (vendor-side signing). Has no Spring or server-runtime dependencies so consumers like cameleer-saas can depend on the minter without inheriting server internals.</description>
Standalone vendor-side tool for producing signed Ed25519 license tokens consumed by `cameleer-server`. The minter is intentionally **not** a runtime or compile-scope dependency of the server — the server only ships with the matching public key and validates tokens via `LicenseValidator`. The private signing key never leaves the vendor's environment.
-`target/cameleer-license-minter-1.0-SNAPSHOT.jar` — plain library JAR (consumable as a Maven `test` dependency or via the `LicenseMinter` API in custom tooling)
-`target/cameleer-license-minter-1.0-SNAPSHOT-cli.jar` — fat CLI JAR with main class `com.cameleer.license.minter.cli.LicenseMinterCli`
## Table of contents
## Audience
## Build
## Public Java API
## CLI usage
## Token format
## LicenseInfo schema
## Limits dictionary
## Generating an Ed25519 key pair
## Worked example
## Security guidance
## Compatibility / runtime separation
---
## Audience
Vendors / SaaS operators issuing licenses to customers who run `cameleer-server`. End-customer operators looking for *how to install* a token should read `docs/license-enforcement.md` instead.
## Build
```bash
# From the repo root
mvn -pl cameleer-license-minter package
```
Two JARs land in `cameleer-license-minter/target/`:
| Artifact | Purpose |
|---|---|
| `cameleer-license-minter-1.0-SNAPSHOT.jar` | Plain library (the `repackage` execution for the main artifact is disabled; see `pom.xml:50-54`). Use this when embedding the minter inside your own tooling or a unit test that needs a fresh signed token. |
| `cameleer-license-minter-1.0-SNAPSHOT-cli.jar` | Fat CLI JAR. Repackaged by Spring Boot's `spring-boot-maven-plugin` with classifier `cli`; main class is `com.cameleer.license.minter.cli.LicenseMinterCli`. |
## Public Java API
`com.cameleer.license.minter.LicenseMinter` is the only entry point for the library. It is a final, stateless utility class:
```java
importcom.cameleer.license.minter.LicenseMinter;
importcom.cameleer.license.LicenseInfo;
LicenseInfoinfo=newLicenseInfo(
java.util.UUID.randomUUID(),
"acme-prod",// tenantId — must match server's CAMELEER_SERVER_TENANT_ID
"Acme Production (Tier B)",// human label, optional
The method is thread-safe; the underlying Jackson `ObjectMapper` is configured once with `ORDER_MAP_ENTRIES_BY_KEYS` so canonical-JSON serialization is deterministic across runs and process boundaries.
`LicenseMinter.mint` will throw `IllegalStateException` if the JCE provider rejects the private key or the payload cannot be serialized.
## CLI usage
The CLI entry point is `com.cameleer.license.minter.cli.LicenseMinterCli`. Run it from the fat JAR produced by the build:
Source of truth: `cameleer-license-minter/src/main/java/com/cameleer/license/minter/cli/LicenseMinterCli.java:26`.
| Flag | Required | Meaning |
|---|---|---|
| `--private-key=<path>` | yes | Path to a PKCS#8-encoded Ed25519 private key. Both PEM (`-----BEGIN PRIVATE KEY-----`) and raw base64 are accepted (`LicenseMinterCli.readEd25519PrivateKey`). |
| `--tenant=<tenantId>` | yes | The exact `tenantId` the server will compare against `CAMELEER_SERVER_TENANT_ID`. Mismatch causes the validator to throw at install / revalidation. |
| `--expires=<YYYY-MM-DD>` | yes | Expiration date interpreted as midnight UTC. The validator considers tokens expired once `now > exp + gracePeriodDays`. |
| `--label=<text>` | no | Human-readable label, surfaced via `GET /api/v1/admin/license` and `/api/v1/admin/license/usage`. |
| `--grace-days=<int>` | no | Number of days the license stays usable after `--expires`. Defaults to `0`. |
| `--max-<limitkey>=<int>` | no, repeatable | Each `--max-foo-bar` flag becomes the limit key `max_foo_bar`. See the limits dictionary below. Unknown keys are accepted by the minter (the server side ignores keys it does not understand and falls through to defaults). |
| `--output=<path>` | no | Write the token to a file. When omitted, the token is printed to stdout. On `--verify` failure the file is deleted. |
| `--public-key=<path>` | no, required for `--verify` | Path to the matching base64 X.509 SPKI public key file (one line, no PEM markers). |
| `--verify` | no | After minting, parse + signature-check the token using `--public-key` and `--tenant`. Exits non-zero if verification fails. |
Exit codes: `0` on success, `1` on minting / IO failure, `2` on argument validation failure, `3` on `--verify` failure.
## Token format
A token is the concatenation of two **standard** base64 segments joined by a literal `.`:
- The canonical JSON payload is produced by `LicenseMinter.canonicalPayload(...)` with keys sorted lexicographically and `limits` rendered as a sorted object. This makes the byte sequence deterministic given a fixed `LicenseInfo`.
- The signature is computed with `Signature.getInstance("Ed25519")` over the canonical payload bytes (not over the base64-encoded form).
- Encoding is `Base64.getEncoder()` (RFC 4648 §4 — *not* base64url). The validator decodes with the matching `Base64.getDecoder()`.
`LicenseValidator.validate(...)` (`cameleer-license-api/src/main/java/com/cameleer/license/LicenseValidator.java:42`) splits on the first `.`, decodes both halves, verifies the signature, then deserializes the payload.
| `licenseId` | `UUID` | yes | Stable identifier for this token. The server's audit trail records install/replace transitions by license id; renewals must use a fresh UUID so audit history is non-ambiguous. |
| `tenantId` | `String` | yes | Must equal the server's `CAMELEER_SERVER_TENANT_ID`. The validator throws `IllegalArgumentException` on mismatch. Blank values are rejected by the canonical record constructor. |
| `label` | `String` | no | Free-form human label. Surfaced on the admin/usage endpoints and the operator UI. Has no enforcement semantics. |
| `limits` | `Map<String,Integer>` | yes (may be empty) | License-specific overrides. Any key that appears here is unioned over `DefaultTierLimits.DEFAULTS` to form the effective caps in `ACTIVE` / `GRACE` states. Keys not present fall through to defaults. |
| `issuedAt` | `Instant` (epoch seconds in JSON `iat`) | yes | Stamped by the minter; not currently consulted by the validator beyond informational logging. |
| `expiresAt` | `Instant` (epoch seconds in JSON `exp`) | yes | The validator throws if `now > expiresAt + gracePeriodDays * 86400` at install or revalidation. |
| `gracePeriodDays` | `int` | yes (>= 0) | Window after `expiresAt` during which the gate transitions to `GRACE` (license still grants its caps) before flipping to `EXPIRED`. Negative values are rejected at construction. |
## Limits dictionary
Canonical key set: `cameleer-license-api/src/main/java/com/cameleer/license/DefaultTierLimits.java`. Any key not listed here is silently ignored by the server's `LicenseGate.getEffectiveLimits()`.
| CLI flag | Key | Default | What the server enforces |
A one-liner using the JDK's `keytool` is **not** sufficient — `keytool` cannot produce raw Ed25519 PKCS#8 in a directly-usable shape for our reader. Generating via the API above (or `openssl genpkey -algorithm ed25519`) is the supported path.
For boot-time installation (preferred for SaaS-managed deployments), set `CAMELEER_SERVER_LICENSE_TOKEN` instead of POSTing — see `docs/license-enforcement.md`.
## Security guidance
- **The Ed25519 private key is the trust root.** Anyone who holds it can mint licenses for any tenant. Treat it like a code-signing key.
- **Storage.** Production private keys belong in an HSM, KMS (e.g. AWS KMS / GCP KMS with non-exportable signing), or a sealed Vault transit backend. A sealed file on a laptop is acceptable for low-volume / pre-production minting only and should never be committed to git or shared via chat.
- **Rotation.** Rotation is destructive: every customer running with the *old* public key will reject all new tokens signed with the *new* private key. The pragmatic procedure is:
1. Generate the new keypair.
2. Distribute the new public key (`CAMELEER_SERVER_LICENSE_PUBLICKEY`) to every tenant's server config.
3. Once tenants confirm they are running with the new public key, re-mint and re-issue every active license under the new key.
4. Decommission the old private key.
Practical revocation flows through expiry — keep license terms short enough (12 months or less) that planned rotations stay aligned with renewal cadence.
- **Auditing.** The server records every install/replace/reject under `AuditCategory.LICENSE`. The minter itself does not write audit rows; if you need a vendor-side audit trail of mint operations, wrap `LicenseMinter.mint(...)` in your own ticketing pipeline.
- **Never commit private keys.** `.gitignore` does not block them by name — use a `secrets/` directory excluded by your repository's policy, or store them entirely outside the working tree.
## Compatibility / runtime separation
The minter is intentionally absent from `cameleer-server-app`'s production classpath. To verify after a build:
# expected: empty output (or, in development branches, a single line scoped 'test')
```
`cameleer-license-minter/pom.xml` depends on `cameleer-license-api` for the pure license contract types (`LicenseInfo`, `LicenseValidator`) used by mint + `--verify`. It deliberately does **not** depend on `cameleer-server-core`, so consumers of the minter (e.g. `cameleer-saas`) do not inherit server-runtime types onto their classpath. The server app intentionally does not depend on the minter — vendors mint outside the customer-deployed runtime, and a compromised customer cannot leverage server code to forge tokens.
description="When to apply: 'live' (default) saves and pushes CONFIG_UPDATE to live agents immediately; 'staged' saves without pushing — the next successful deploy applies it.")
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.