[P3] Comparative & historical analysis #115

Open
opened 2026-04-01 22:56:09 +02:00 by claude · 1 comment
Owner

Parent Epic

#100

Problem

"How did this route perform last week vs. this week?" — there's no comparison view, no regression detection, and no way to overlay historical data. Competing tools like Datadog and Grafana offer week-over-week comparisons that help teams identify regressions after deployments.

Proposed Solution

1. Time Comparison on Dashboard

Add a "Compare" toggle on the dashboard that overlays a previous time period:

┌─ THROUGHPUT BY ROUTE ──────────── Compare: [Previous period ▾] ─┐
│                                                                  │
│  ───── current (solid line)                                      │
│  - - - previous period (dashed, lighter color)                   │
│                                                                  │
│  Tooltip: "route3: 45 msg/s (was 38 msg/s, +18%)"              │
└──────────────────────────────────────────────────────────────────┘

Compare options:

  • Previous period (same duration, prior window)
  • Same time last week
  • Same time last month
  • Custom range

2. Deployment Markers

Show vertical lines on timeseries charts when a new agent version was deployed:

│  ▼ v1.2.3 deployed                                              │
│  │                                                               │
│  ──────┤─────────────────────────────                            │
│        │                                                         │
│  (latency increased after deploy)                                │

The agent already reports its version — detect version changes and mark them.

3. Route Performance Diff

A dedicated comparison view:

┌─ Performance Comparison ────────────────────────────────┐
│  Route: complex-fulfillment                             │
│  This week vs. Last week                                │
│                                                         │
│  Metric          This Week    Last Week    Change       │
│  Throughput      12.3 msg/s   11.8 msg/s   +4.2%  ✓   │
│  Success Rate    98.5%        99.1%         -0.6%  ⚠   │
│  P99 Latency     450ms        380ms         +18%   ⚠   │
│  Error Count     23           12            +92%   ✗   │
└─────────────────────────────────────────────────────────┘

Acceptance Criteria

  • Dashboard charts support previous-period overlay
  • Deployment markers on timeseries charts
  • Performance diff view comparing two time periods
  • Change indicators (%, direction, severity)

Notes

This is a differentiation feature that elevates Cameleer from "monitoring" to "analytics." Not required for launch but would strengthen the competitive story significantly.

## Parent Epic #100 ## Problem "How did this route perform last week vs. this week?" — there's no comparison view, no regression detection, and no way to overlay historical data. Competing tools like Datadog and Grafana offer week-over-week comparisons that help teams identify regressions after deployments. ## Proposed Solution ### 1. Time Comparison on Dashboard Add a "Compare" toggle on the dashboard that overlays a previous time period: ``` ┌─ THROUGHPUT BY ROUTE ──────────── Compare: [Previous period ▾] ─┐ │ │ │ ───── current (solid line) │ │ - - - previous period (dashed, lighter color) │ │ │ │ Tooltip: "route3: 45 msg/s (was 38 msg/s, +18%)" │ └──────────────────────────────────────────────────────────────────┘ ``` Compare options: - Previous period (same duration, prior window) - Same time last week - Same time last month - Custom range ### 2. Deployment Markers Show vertical lines on timeseries charts when a new agent version was deployed: ``` │ ▼ v1.2.3 deployed │ │ │ │ │ ──────┤───────────────────────────── │ │ │ │ │ (latency increased after deploy) │ ``` The agent already reports its version — detect version changes and mark them. ### 3. Route Performance Diff A dedicated comparison view: ``` ┌─ Performance Comparison ────────────────────────────────┐ │ Route: complex-fulfillment │ │ This week vs. Last week │ │ │ │ Metric This Week Last Week Change │ │ Throughput 12.3 msg/s 11.8 msg/s +4.2% ✓ │ │ Success Rate 98.5% 99.1% -0.6% ⚠ │ │ P99 Latency 450ms 380ms +18% ⚠ │ │ Error Count 23 12 +92% ✗ │ └─────────────────────────────────────────────────────────┘ ``` ## Acceptance Criteria - [ ] Dashboard charts support previous-period overlay - [ ] Deployment markers on timeseries charts - [ ] Performance diff view comparing two time periods - [ ] Change indicators (%, direction, severity) ## Notes This is a differentiation feature that elevates Cameleer from "monitoring" to "analytics." Not required for launch but would strengthen the competitive story significantly.
claude added the featureux labels 2026-04-01 22:56:09 +02:00
Author
Owner

Design Specification

Time comparison toggle on dashboard charts: [Off] [Previous period] [Last week] [Custom]. Comparison data rendered as dashed lines with 40% opacity (line charts) or 15% fill (area charts).

Deployment markers: Vertical dashed reference lines on charts when agent version changes, with version label. Data from agent_metrics version transitions.

Performance diff table (when comparison active):

Metric        | Current  | Previous | Change
Throughput    | 1,234/m  | 1,180/m  | +4.6% ↑
P99 Latency   | 142ms    | 128ms    | +10.9% ↑
Error Count   | 22       | 11       | +100% ↑

Green for improvements, red for regressions, grey for <1% changes.

Backend: Extend timeseries endpoints with compare=previous|week param returning comparisonBuckets[]. New /agents/deployments endpoint for version transitions.

## Design Specification **Time comparison toggle** on dashboard charts: `[Off] [Previous period] [Last week] [Custom]`. Comparison data rendered as dashed lines with 40% opacity (line charts) or 15% fill (area charts). **Deployment markers**: Vertical dashed reference lines on charts when agent version changes, with version label. Data from agent_metrics version transitions. **Performance diff table** (when comparison active): ``` Metric | Current | Previous | Change Throughput | 1,234/m | 1,180/m | +4.6% ↑ P99 Latency | 142ms | 128ms | +10.9% ↑ Error Count | 22 | 11 | +100% ↑ ``` Green for improvements, red for regressions, grey for <1% changes. Backend: Extend timeseries endpoints with `compare=previous|week` param returning `comparisonBuckets[]`. New `/agents/deployments` endpoint for version transitions.
Sign in to join this conversation.