logging-best-practices

Canonical log lines are single, context-rich JSON events emitted per request per service. They replace scattered logs with high-cardinality data, enabling fast debugging and powerful filtering across distributed microservice architectures.

Install a JSON logger at startup, implement middleware to collect and emit wide events, and wrap handlers to enrich them with business context. Capture environment data once to ensure a consistent schema across all emitted events.

Wide-event logging applies to any backend service, API, or microservice architecture. It requires no specific platform, relying instead on middleware to enforce a single JSON schema and capture environment-aware context.

The best way is using the wide events pattern to include user, request, session, and business context. Middleware automatically enriches these JSON logs with environment data like version and region for consistent queryability.

Middleware automatically captures environment data such as version, region, and deploy ID at startup. It then wraps handlers to append high-cardinality business context, emitting a single JSON-formatted event per request.

Canonical log lines support distinguishing info and error events within the single JSON logger schema. This separation maintains consistent queryability while allowing teams to filter specific event types during analysis.