Observability

Production services need structured logs, traces, and metrics.

Logging

Use request-scoped IDs and structured JSON logs for correlation.

Recommended fields:

• request id
• route pattern
• HTTP method/status
• duration
• user id (if authenticated)

Tracing

Instrument critical paths:

• DB queries and transactions
• outbound HTTP calls
• queue enqueue/dequeue
• websocket room operations

Use span boundaries around handlers and service-layer operations.

Metrics

Track at minimum:

• request rate
• latency percentiles
• error rate by route
• queue depth and retry count
• cache hit/miss ratio

Practical Deployment Notes

• Keep high-cardinality labels out of metric keys.
• Sample traces in high-throughput environments.
• Tie request IDs across logs and traces.