Observability & SRE
# Metrics, logs, and traces; SLOs and error budgets; incident response — running systems as an engineering discipline.
❯ cat documents/
-
Observability & SRE practice
Concept
2026-08-08
Metrics, logs, and traces; SLOs and error budgets; incident response — running systems as an engineering discipline.
-
Structured logging & MDC
Concept
2026-08-11
Logs as machine-parseable events with per-request context — JSON output, level discipline, what belongs in a log vs a metric vs a span, and the MDC idiom across SLF4J, slog, tracing, and AsyncLocalStorage.
-
Correlation IDs & context propagation
Concept
2026-08-11
Extract-or-generate an id at the edge, carry it in a request context, stamp it on logs, responses, and outgoing calls — and where business correlation ids end and W3C traceparent/baggage begin.
-
OpenTelemetry
Concept
2026-08-11
The vendor-neutral telemetry standard — the API/SDK/Collector split, the three signals with their data models, semantic conventions, OTLP and the OTEL_* env vars, and auto vs manual instrumentation per platform.
-
Health checks: liveness, readiness, startup
Concept
2026-08-11
Designing probe endpoints from the probed side — liveness without dependency checks, readiness as the traffic gate, startup probes for slow boots — and the endpoint catalogue across Actuator, SmallRye, Micronaut, and hand-rolled stacks.