feat: progressive drill-down dashboard with RED metrics and SLA compliance (#94)

Three-level dashboard driven by sidebar selection:
- L1 (no selection): all-apps overview with health table, per-app charts
- L2 (app selected): route performance table, error velocity, top errors
- L3 (route selected): processor table, latency heatmap data, bottleneck KPI

Backend: 3 new endpoints (timeseries/by-app, timeseries/by-route, errors/top),
per-app SLA settings (app_settings table, V12 migration), exact SLA compliance
from executions hypertable, error velocity with acceleration detection.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

cameleer3-server-core/src/main/java/com/cameleer3/server/core/admin/AppSettings.java

@@ -0,0 +1,19 @@
+package com.cameleer3.server.core.admin;
+
+import java.time.Instant;
+
+public record AppSettings(
+        String appId,
+        int slaThresholdMs,
+        double healthErrorWarn,
+        double healthErrorCrit,
+        double healthSlaWarn,
+        double healthSlaCrit,
+        Instant createdAt,
+        Instant updatedAt) {
+
+    public static AppSettings defaults(String appId) {
+        Instant now = Instant.now();
+        return new AppSettings(appId, 300, 1.0, 5.0, 99.0, 95.0, now, now);
+    }
+}

cameleer3-server-core/src/main/java/com/cameleer3/server/core/admin/AppSettingsRepository.java

@@ -0,0 +1,11 @@
+package com.cameleer3.server.core.admin;
+
+import java.util.List;
+import java.util.Optional;
+
+public interface AppSettingsRepository {
+    Optional<AppSettings> findByAppId(String appId);
+    List<AppSettings> findAll();
+    AppSettings save(AppSettings settings);
+    void delete(String appId);
+}

cameleer3-server-core/src/main/java/com/cameleer3/server/core/search/ExecutionStats.java

@@ -14,4 +14,23 @@ public record ExecutionStats(
         long prevTotalCount,
         long prevFailedCount,
         long prevAvgDurationMs,
-        long prevP99LatencyMs) {}
+        long prevP99LatencyMs,
+        double slaCompliance) {
+
+    /** Constructor without SLA compliance (backward-compatible, sets to -1). */
+    public ExecutionStats(long totalCount, long failedCount, long avgDurationMs,
+                          long p99LatencyMs, long activeCount, long totalToday,
+                          long prevTotalCount, long prevFailedCount,
+                          long prevAvgDurationMs, long prevP99LatencyMs) {
+        this(totalCount, failedCount, avgDurationMs, p99LatencyMs, activeCount,
+                totalToday, prevTotalCount, prevFailedCount, prevAvgDurationMs,
+                prevP99LatencyMs, -1.0);
+    }
+
+    /** Return a copy with the given SLA compliance value. */
+    public ExecutionStats withSlaCompliance(double slaCompliance) {
+        return new ExecutionStats(totalCount, failedCount, avgDurationMs, p99LatencyMs,
+                activeCount, totalToday, prevTotalCount, prevFailedCount,
+                prevAvgDurationMs, prevP99LatencyMs, slaCompliance);
+    }
+}

cameleer3-server-core/src/main/java/com/cameleer3/server/core/search/SearchService.java

@@ -5,6 +5,7 @@ import com.cameleer3.server.core.storage.StatsStore;
 
 import java.time.Instant;
 import java.util.List;
+import java.util.Map;
 
 public class SearchService {
 
@@ -48,4 +49,38 @@ public class SearchService {
                                        String routeId, List<String> agentIds) {
         return statsStore.timeseriesForRoute(from, to, bucketCount, routeId, agentIds);
     }
+
+    // ── Dashboard-specific queries ────────────────────────────────────────
+
+    public Map<String, StatsTimeseries> timeseriesGroupedByApp(Instant from, Instant to, int bucketCount) {
+        return statsStore.timeseriesGroupedByApp(from, to, bucketCount);
+    }
+
+    public Map<String, StatsTimeseries> timeseriesGroupedByRoute(Instant from, Instant to,
+                                                                   int bucketCount, String applicationName) {
+        return statsStore.timeseriesGroupedByRoute(from, to, bucketCount, applicationName);
+    }
+
+    public double slaCompliance(Instant from, Instant to, int thresholdMs,
+                                String applicationName, String routeId) {
+        return statsStore.slaCompliance(from, to, thresholdMs, applicationName, routeId);
+    }
+
+    public Map<String, long[]> slaCountsByApp(Instant from, Instant to, int defaultThresholdMs) {
+        return statsStore.slaCountsByApp(from, to, defaultThresholdMs);
+    }
+
+    public Map<String, long[]> slaCountsByRoute(Instant from, Instant to,
+                                                  String applicationName, int thresholdMs) {
+        return statsStore.slaCountsByRoute(from, to, applicationName, thresholdMs);
+    }
+
+    public List<TopError> topErrors(Instant from, Instant to, String applicationName,
+                                     String routeId, int limit) {
+        return statsStore.topErrors(from, to, applicationName, routeId, limit);
+    }
+
+    public int activeErrorTypes(Instant from, Instant to, String applicationName) {
+        return statsStore.activeErrorTypes(from, to, applicationName);
+    }
 }

cameleer3-server-core/src/main/java/com/cameleer3/server/core/search/TopError.java

@@ -0,0 +1,12 @@
+package com.cameleer3.server.core.search;
+
+import java.time.Instant;
+
+public record TopError(
+        String errorType,
+        String routeId,
+        String processorId,
+        long count,
+        double velocity,
+        String trend,
+        Instant lastSeen) {}

cameleer3-server-core/src/main/java/com/cameleer3/server/core/storage/StatsStore.java

@@ -2,9 +2,11 @@ package com.cameleer3.server.core.storage;
 
 import com.cameleer3.server.core.search.ExecutionStats;
 import com.cameleer3.server.core.search.StatsTimeseries;
+import com.cameleer3.server.core.search.TopError;
 
 import java.time.Instant;
 import java.util.List;
+import java.util.Map;
 
 public interface StatsStore {
 
@@ -33,4 +35,29 @@ public interface StatsStore {
     // Per-processor timeseries
     StatsTimeseries timeseriesForProcessor(Instant from, Instant to, int bucketCount,
                                             String routeId, String processorType);
+
+    // Grouped timeseries by application (for L1 dashboard charts)
+    Map<String, StatsTimeseries> timeseriesGroupedByApp(Instant from, Instant to, int bucketCount);
+
+    // Grouped timeseries by route within an application (for L2 dashboard charts)
+    Map<String, StatsTimeseries> timeseriesGroupedByRoute(Instant from, Instant to, int bucketCount,
+                                                           String applicationName);
+
+    // SLA compliance: % of completed exchanges with duration <= thresholdMs
+    double slaCompliance(Instant from, Instant to, int thresholdMs,
+                         String applicationName, String routeId);
+
+    // Batch SLA counts by app: {appId -> [compliant, total]}
+    Map<String, long[]> slaCountsByApp(Instant from, Instant to, int defaultThresholdMs);
+
+    // Batch SLA counts by route within an app: {routeId -> [compliant, total]}
+    Map<String, long[]> slaCountsByRoute(Instant from, Instant to, String applicationName,
+                                          int thresholdMs);
+
+    // Top N errors with velocity trend
+    List<TopError> topErrors(Instant from, Instant to, String applicationName,
+                             String routeId, int limit);
+
+    // Count of distinct error types in window
+    int activeErrorTypes(Instant from, Instant to, String applicationName);
 }