Correlation ID
Also known as:
Request ID
Message ID
Trace ID
Definition
A correlation id is a unique value attached to a request so its response can be matched back to it later, even when many requests are in flight and replies arrive out of order. The sender keeps a map of correlation id to pending request, and when a response echoes the same id, the sender knows exactly which request it belongs to. It is what makes asynchronous messaging, pipelining, and a request waiting list work.
Key Takeaways
- It ties an asynchronous response back to the request that caused it, without relying on arrival order.
- Every response must echo the same id the request carried, so the receiver can look it up.
- It is the key that a pending-requests map or request waiting list is indexed on.
- The same id, propagated across services, becomes a trace id that lets you follow one request through a whole system.
How It Works
- The sender generates a unique id and stamps it on the outgoing request.
- It stores the id in a map that points to the waiting caller, future, or callback.
- The receiver copies the id onto its response.
- When the response arrives, the sender reads the id, finds the matching entry, and completes it.
Where It Is Used
- gRPC and HTTP/2 use per-stream ids so multiplexed responses can arrive in any order.
- Message brokers and RPC frameworks stamp a correlation id on each message for request-reply.
- Distributed tracing tools like OpenTelemetry propagate a trace id, a correlation id spanning many services.