Skip to content

feat(observability): structured logging, metrics, tracing #10

@MarkCesium

Description

@MarkCesium

Summary

backend/auth/src/main.py:27-31 uses logging.basicConfig with a text format. No correlation across services, no metrics endpoint, no traces.

Changes

  • Replace logging setup with structlog emitting JSON; bind correlation_id, user_id (when known), grpc_method.
  • gRPC server interceptor: read x-correlation-id from metadata, generate if missing, bind to structlog.contextvars.
  • Add a sidecar HTTP server on AUTH__METRICS_PORT exposing /metrics via prometheus-client.
  • Instrument gRPC server + SQLAlchemy + Redis with OpenTelemetry auto-instrumentation; export OTLP to the collector from infra#18.
  • Histograms: auth_grpc_request_duration_seconds{method}. Counters: login success/failure, OTP sends, revocation hits.

Verification

  • docker compose logs auth emits one JSON object per line.
  • curl auth:9100/metrics returns Prometheus exposition format.
  • A login request from gateway produces a span chain visible in Tempo/Jaeger with the gateway parent span.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions