Repository navigation
feat(logging): log candidate vs reference latency delta - #27
Merged
Merged
Conversation
Response comparison correlation logs showed whether an ingress candidate matched the reference but not whether it was faster or slower, so latency regressions were invisible from the logs alone. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Response comparison correlation logs recorded whether each ingress candidate matched the reference, but not whether the candidate was faster or slower. That left latency regressions invisible from the correlation logs alone.
This records each backend's time to first byte in the ingress capture writer and logs their difference as
dton the "Response comparison completed" line, for exampledt=-143ms. The value is the candidate's latency minus the reference's: negative when the candidate was faster, positive when it was slower. It is omitted when either side was not timed.Time to first byte is used rather than full response time so the reference, which streams its body to the real client, is not penalised by a slow downstream reader while the candidate drains to a discard sink.
The egress correlation is left unchanged: there the candidate replays the reference's recorded response instead of making its own dependency call, so there is no independent candidate latency to compare.
🤖 Generated with Claude Code