Repository navigation
Give inference requests an X-Request-ID - #442
Merged
Merged
Conversation
The inference app never set a request id, so its logs and error responses had none and its calls to the gateway could not be matched to a request. Add a plain ASGI request-id middleware to common (it leaves streamed responses untouched), install it on the inference app, and forward the id in ApiGatewayClient. A caller's id is kept only if it is short and plain. Signed-off-by: Priyanshu-u07 <connect.priyanshu8271@gmail.com>
This was referenced Oct 26, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The inference app never set a request id, so
request_id_ctxwas always empty there: its JSON logs and error responses carried no id, and its calls to the gateway could not be matched to a request.RequestIDMiddlewaretocommon/middleware.py. It's plain ASGI rather thanBaseHTTPMiddleware, so streamed responses pass through untouched. It keeps a caller'sX-Request-IDif it's short and plain, otherwise generates one, and sets it inrequest_id_ctx,request.state.request_idand the response header.ApiGatewayClientso gateway logs can be matched to the inference request.Tests:
test_request_id.py(8 tests: new id, kept id, rejected ids, streaming, the real app, forwarding to the gateway). Full suite with the CI ignore list: 3,061 passed.