Question

A company runs a microservices application instrumented with AWS X-Ray. During a recent incident, the on-call team spent significant time manually clicking through individual traces to find requests that failed for a specific enterprise customer (identified by a customer ID passed in each request). The team wants a repeatable way to quickly isolate only the faulted traces belonging to a particular customer ID so they can diagnose issues faster in future incidents. What should the team implement?

Accepted Answer

Add the customer ID as an X-Ray annotation in the instrumented segments, then use X-Ray filter expressions to query for faulted traces matching that annotation value.

Answer

Add the customer ID as an X-Ray metadata field in the segments, then use X-Ray filter expressions to search for the customer ID and fault status.

Answer

Increase the X-Ray sampling rate to 100% for all services so that every request for the customer is captured and can be found in the trace list.

Answer

Enable CloudWatch Logs for each service and use Logs Insights to correlate the customer ID with HTTP 5xx status codes across all log groups.