observability-and-instrumentation

Structured logging means emitting JSON objects with a stable event name and machine-readable fields instead of interpolated strings. Attach a correlation ID via a child logger per request, use consistent log levels, and never log secrets or full request bodies.

Use the RED method: rate, errors, and duration for every endpoint and external dependency. Track latency as a histogram so you can query p95/p99 percentiles, and keep label values bounded to small fixed sets like route templates and status classes.

They serve different signals: OpenTelemetry provides vendor-neutral tracing and metrics APIs with auto-instrumentation for HTTP and database clients, while Prometheus is a common metrics backend. The RED/USE and cardinality rules apply identically regardless of backend choice.

Every unique label combination creates a separate time series, so unbounded values like user IDs, raw URLs, or error messages explode cardinality and can crash the metrics backend. High-cardinality lookups belong in logs and traces instead.

Do not use it to diagnose an active incident, which belongs to debugging workflows, or to profile measured slowness, which belongs to performance optimization. This Skill covers writing telemetry alongside features, not operating on it after the fact.