Currently all log output is plain text which makes it difficult to trace a single request across multiple services in production.
Proposal: switch to structured JSON logging and inject a correlation ID at the request boundary. The ID should propagate through all downstream calls via headers.
Benefits:
- Enables log aggregation tools (Datadog, Grafana Loki) to group related events
- Dramatically reduces MTTR for production incidents
- Makes distributed tracing integration straightforward
Suggested library: pino (Node) or zerolog (Go) depending on the service.
Currently all log output is plain text which makes it difficult to trace a single request across multiple services in production.
Proposal: switch to structured JSON logging and inject a correlation ID at the request boundary. The ID should propagate through all downstream calls via headers.
Benefits:
Suggested library: pino (Node) or zerolog (Go) depending on the service.