Clarify use of trace-id in response header
Closed this issue · 3 comments
After feedback from @adriancole it became clear that the use-case for trace-id is not well defined in the spec. Update the wording to make it clear that the caller is not meant to adopt the trace id.
The use case for returning trace ID is when cloud provider returns an internal trace reference. What is the proposed alternative for that use cae?
There is no proposed alternative. The issue is that the use case you stated wasn't clear to everyone. This issue is just to remind me to take a look through the spec and make sure it's clear that this is the intended use case.
My understanding is that the use case is:
- Client sends a request to a managed service in the cloud. Client may or may not include traceparent.
- Cloud provider internally starts another trace, and returns its ID in the response header.
- Client records the response trace ID in its own span/trace, or logs it, so that it can be given to cloud customer support for troubleshooting.
When I say "trace ID", it may also include the span ID, if the cloud provider wants to have more precise reference (perhaps in some messaging / serverless scenarios the reference is not always to the root), but that's a speculation.