Observability¶

Logging¶

• Adopt structured logs with a correlation/request ID propagated across services.
• Include client_id, job_id, user_id (where appropriate, non-PII), and provider.

Metrics (examples)¶

• HTTP: req/sec, p95 latency, error rate (4xx/5xx) per endpoint.
• Jobs: queue depth, time-to-complete, success/failure ratio.
• Redis: command latency, error rate, CPU/memory.
• Docs build: build time, success/failure counts.

Alerts¶

• CPM/External API: 5xx > 2% for 5 min.
• Redis: p95 latency > 100ms for 5 min; error rate > 1%.
• Docs: build workflow failures on main/tags.

Dashboards¶

• API overview, Job lifecycle, Redis health, Deployment status.