When Mule creates a new event, it needs a correlation ID so log entries and processing activity associated with that event can be traced together. Mule first checks whether the event source already provides a correlation ID. For an HTTP Listener, headers such as X-CORRELATION-ID or MULE_CORRELATION_ID can supply that identifier.

MuleSoft explicitly states that if the source message contains a correlation ID, Mule uses it; if the source does not provide one, Mule generates a unique correlation ID. Therefore, option C accurately describes HTTP Listener behavior.

Option B is incorrect because Mule does not unconditionally replace an incoming correlation ID. Option D is also incorrect: MuleSoft specifically documents that Anypoint MQ does not automatically propagate Mule correlationId between publisher and subscriber. If an identifier must traverse Anypoint MQ, it should be included as a user property and then reapplied or used by the receiving application.

Correlation IDs are central to production observability because they let operations teams connect log records and processing stages to the same business execution.

Reference topics: Mule correlation IDs; HTTP Listener; X-CORRELATION-ID; distributed tracing; event correlation.

Official documentation: https://docs.mulesoft.com/mule-runtime/latest/correlation-id

Official documentation: https://docs.mulesoft.com/mq/mq-faq

===============================================================