feat(pipeline): coalesce multi-line stack traces into one event#1
Merged
Conversation
A multi-line stack trace previously exploded into one template per frame — a Python FastAPI traceback became ~40 "novel" templates in M4 live testing, which is exactly the noise logscry exists to suppress. Add a coalescer stage upstream of templating: a standalone goroutine folds continuation lines (indentation, frame markers, language cues for Python, Java, and Go) into the preceding logical line, so a traceback reaches templating as ONE model.LogLine and dedups as a single event. The heuristic is conservative — unrelated single-line logs stay separate — and gated leniency (blank lines, exception summaries, Go frames, the SIGSEGV "[signal ...]" line) only applies once a multi-line block is confirmed active. Bounded latency: a buffered event is flushed on the next header, after an idle timeout, or when input closes. The timeout is configurable via group.timeout / --group-timeout (default 200ms; 0 disables). Coalescer state is confined to the single goroutine, so the no-lock concurrency model (RDI §3) is unchanged and go test -race stays green. Masking order and rules are untouched. 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A multi-line stack trace previously exploded into one template per frame — a Python FastAPI traceback became ~40 "novel" templates in M4 live testing, which is exactly the noise logscry exists to suppress.
Add a coalescer stage upstream of templating: a standalone goroutine folds continuation lines (indentation, frame markers, language cues for Python, Java, and Go) into the preceding logical line, so a traceback reaches templating as ONE model.LogLine and dedups as a single event. The heuristic is conservative — unrelated single-line logs stay separate — and gated leniency (blank lines, exception summaries, Go frames, the SIGSEGV "[signal ...]" line) only applies once a multi-line block is confirmed active.
Bounded latency: a buffered event is flushed on the next header, after an idle timeout, or when input closes. The timeout is configurable via group.timeout / --group-timeout (default 200ms; 0 disables). Coalescer state is confined to the single goroutine, so the no-lock concurrency model (RDI §3) is unchanged and go test -race stays green. Masking order and rules are untouched.