Add structured request logs and correlation IDs

Distributed calls need a common trace while individual hops keep their own request identities. This lets an operator follow one task without logging its entire payload.

Propagate fixture context

The integration lab maps incoming, outbound and response identifiers.

from integration_cases import request_log_case

result = request_log_case()
assert result["trace_propagated"] is True
assert result["outbound"]["parent_request_id"] == "R-in"
assert result["response"]["request_id"] == "R-in"
assert result["raw_body_logged"] is False

Accept only valid bounded incoming IDs or generate a new one. Put the request/trace in logging context, outbound headers, queue metadata and error responses. Preserve parent-child relationships rather than reusing one hop ID everywhere.

Log method/route template, status, duration, release and safe user/tenant pseudonym. Avoid query strings, authorization and bodies by default. High-cardinality content belongs in neither metric labels nor broad logs.

The FDE for Freshers course uses correlation in deployment monitoring and debugging.

Exercise

Trace one local request through an API, worker and fake dependency. Query the JSON log by trace ID and prove a seeded token is absent.

Continue learning

This article is part of the FDE integration and deployment foundations sequence. Use the neighbouring tasks when you need the prerequisite or the next application.

• Review the prerequisite or neighbouring task in Separate liveness and readiness checks.
• Continue with Measure API latency before optimizing it.

Reference: OpenTelemetry trace conventions.