agent-observability

Observability instrumentation requires Sentry or Supabase integration alongside structured logging with a correlation_id. You collect logs, metrics, and traces across edge layers, APIs, cron jobs, and data pipelines to detect errors and latency hotspots.

Defining SLI and SLO thresholds requires measuring latency and error rates to configure actionable alerts. This approach prevents noise by ensuring alerts only fire when performance degradation threatens established service level objectives.

Yes, Sentry and Supabase integration supports structured logging and alert routing. You use a runtime_logs schema with a correlation_id to aggregate logs and route alerts to appropriate responders before customers are impacted.

A correlation_id in logging and metrics traces requests across edge, API, and pipeline components. This guarded logging practice validates data governance and links logs to metrics, reducing entropy and making incident root causes identifiable.

Running a blameless post-mortem after incident response involves documenting the event using operational dashboards and runbooks. This governance practice analyzes collected traces and metrics to prevent future recurrence without attributing individual fault.