logging-best-practices

Structured wide-event logging emits a single, context-rich JSON event per request to simplify cross-service querying. It solves debugging complexity by consolidating business, environment, and timing data into one canonical log line for fast insights.

To implement wide-event logging, use middleware to initialize a single globally shared logger at request start, enrich it with business and environment context, and emit the JSON event in a finally block to ensure consistent lifecycle management.

Propagate request IDs across services by enforcing a consistent JSON schema and passing the request ID through middleware. This ensures every wide event carries the shared identifier, enabling accurate cross-service tracing and observability.

Canonical log lines should include high cardinality and high dimensionality context like user IDs, request IDs, business data such as cart value, and environment details like version and region to enable flexible, fast analyses.

Yes, structured wide-event logging requires a single globally shared logger instance per service. This ensures consistent JSON formatting and schema enforcement across all middleware-enriched events emitted by the application.