hatch3r-observability-verify

To verify OpenTelemetry tracing coverage, validate that every request path has root spans and outbound calls have child spans with required HTTP and DB semantic attributes. This ensures end-to-end observability for new API routes before they reach production.

Trace_id correlation in structured logging ensures every request-scoped JSON log line includes trace identifiers for click-through debugging. It is needed to connect logs with traces, preventing blind spots during incident investigation.

To enforce SLOs and multi-window burn-rate alerting as code, validate committed SLO definitions, review burn-rate alerts, and ensure alerts-as-code include runbook links. This maintains reliability standards across web services and queue consumers.

Yes, observability verification works with queue consumers and LLM-invoking features by applying tracing, structured logging, and RED/USE metrics gates. It also validates optional GenAI semantic convention instrumentation when applicable.

The best way to prevent deploying services without metrics and error tracking is to apply an observability gate that checks RED/USE instrumentation and error-tracker integration. This validates production readiness before marking features complete.

To control metric cardinality budgets with exemplars, validate metric instrumentation against defined cardinality limits and ensure exemplars are attached. This prevents unbounded metric cardinality from causing storage and query performance issues.