fix(probe): cancel streaming response to release concurrency slot#442
Merged
Conversation
Chat probes use stream: true to measure realistic TTFT/TPS, but the ProbeService never consumed or cancelled the returned ReadableStream. The dispatcher wraps streams so doRelease() only fires when the stream is consumed, cancelled, or errored. Since the probe only reads response.plexus metadata, the stream was silently abandoned, leaking one concurrency slot per chat probe. Over time this exhausted maxConcurrency and deadlocked the provider/model. Cancel the stream after dispatch with error handling (logger.warn on failure, since the slot is already released by the wrapper).
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fix an issue where
runProbenever releases its concurrency slot.Steps to reproduce:
Root cause:
Chat probe uses
stream: trueto measure TTFT/TPS, butProbeServicenever consumed or cancelled the returnedReadableStream. SincedoRelease()only fires when a stream is consumed, cancelled, or errored, the probe will permanently leak 1 concurrency slot per each run.Fix:
Make sure to call
response.stream.cancel()after running probes.