System Health & Validation
Canary Tests & Component Monitoring — Prove the factory works
Build Specification: System Health & Validation Dashboard
Specification Source: Hilbert Factory Section 19 (System Validation & Canary Test Protocol) + Dashboard Spec View 7
Panel 7.1 — Component Health Grid
Chart Type: Status Grid (Custom HTML/CSS)
Data Source: GET /api/health/components — returns all 10 component statuses
Refresh Rate: REALTIME (every 60 seconds, manual refresh button forces immediate)
Display: Grid of 10 components. Each: name, status indicator (green ✓ / yellow △ / red ✗), response time (ms), uptime % over 24 hours.
Components: PostgreSQL, Knowledge Graph, Template Registry API, Prompt Registry, Artifact Storage, Claude API, Build Orchestrator, Prompt Execution Engine, Builder Agent Pool, Verifier Agent Pool
Interaction: Click any component → 24-hour health history chart + error log
Panel 7.2 — Canary Test History
Chart Type: Data Table + Gauge
Data Source: GET /api/canary/history — returns recent canary runs
Refresh Rate: PERIODIC (every 5 minutes)
Display: Most recent run at top with large PASS/FAIL indicator. Table: canary_id, trigger (SCHEDULED/MANUAL), result, failing step if FAIL, latency, cost, timestamp. Pass rate gauge: should be 100% over 30 days.
Panel 7.3 — Run Canary Test Button
Chart Type: Action Button + Progress Display (Custom HTML/JS)
Trigger: Button click → POST /api/canary/run
Display: Prominent button: “Run Canary Test” with canary icon. On click: real-time progress — Readiness ✓ → Queued ✓ → Building... → Verifying... → Complete ✓ (or ✗ at failing step). Result: green banner “PASSED” or red banner “FAILED at [step]”.
Cooldown: 5 minutes between manual runs. Button greyed out during cooldown with countdown timer.
Panel 7.4 — Run Health Check Button
Chart Type: Action Button (Custom HTML/JS)
Trigger: Button click → POST /api/health/check-all
Display: Button: “System Health Check” with heartbeat icon. On click: runs all 10 component checks immediately. Results populate Panel 7.1. Summary: “10/10 healthy” or “X/10 healthy — [failing components]”.
Panel 7.5 — Factory Commissioning Status
Chart Type: Metric Card
Data Source: GET /api/validation/commissioning — returns commissioning proof object reference
Refresh Rate: SESSION
Display: “COMMISSIONED: [date]” with green badge, or “NOT COMMISSIONED” with amber badge. Link to Factory Commissioning Proof Object. Last pre-production validation date and result.