[reference] feat(kafka-clients): toolkit-generated instrumentation (blind test attempt4) - #12361
Draft
jordan-wong wants to merge 2 commits into
Draft
Conversation
… blind test attempt4) Blind regeneration of kafka-clients-0.11/-3.8 + kafka-common/-connect/-streams against the current apm-instrumentation-toolkit, testing whether an explicit per-item-span skill rule (dd-trace-java PR #12293) changes agent behavior for batch-consume operations. It did not: the agent reproduced the correct per-message-span pattern (TracingIterable/TracingIterator wrapping poll()/records()/iterator() returns) by copying master's existing implementation directly, without reading the new skill guidance. [reference] / [DO NOT MERGE] -- this is a research artifact for reviewer feedback, not a merge candidate. See docs/eval-research/runs/kafka-clients/attempt4/outcome.md in the apm-instrumentation-toolkit repo for the full analysis. Final toolkit review status: approved=False, todos_remaining=18 (mostly low-priority comment/naming-hygiene items and one spurious version-alias finding, not architectural defects).
jordan-wong
added
tag: ai generated
This comment has been minimized.
This comment has been minimized.
Contributor
🟢 Java Benchmark SLOs — All performance SLOs passed
PR vs. master results
Commit: Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion. |
Kafka / producer-benchmarkParameters
See matching parameters
SummaryFound 0 performance improvements and 0 performance regressions! Performance is the same for 3 metrics, 0 unstable metrics. See unchanged results
|
Kafka / consumer-benchmarkParameters
See matching parameters
SummaryFound 0 performance improvements and 0 performance regressions! Performance is the same for 3 metrics, 0 unstable metrics. See unchanged results
|
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.
What Does This Do
[DO NOT MERGE] — this is a research reference artifact, not a merge candidate.
Blind regeneration of the kafka-clients-0.11/-3.8 + kafka-common/kafka-connect-0.11/kafka-streams-0.11/-1.0 family against the current apm-instrumentation-toolkit, with the family fully deleted from HEAD before generation (blind test protocol).
This run specifically tested whether an explicit skill rule added this cycle (
.agents/skills/apm-integrations/references/advice-class.md— "Batch-consume operations: one span per item, not one span for the whole batch", see #12293) changes agent behavior for the batch-consume/per-item-span pattern.Motivation
Result: the agent reproduced the correct per-message-span pattern (
TracingIterable/TracingIterator/TracingListwrappingpoll()/records()/iterator()returns, matching master's real implementation) — but by copying master's existing helper classes directly viagit show, not by reading the new skill guidance (zero transcript references toadvice-class.md). A follow-up fix force-reading that file at code-gen has been opened as apm-instrumentation-toolkit PR #622; a re-run against that fix is in progress (attempt5).Full analysis:
docs/eval-research/runs/kafka-clients/attempt4/outcome.mdin the apm-instrumentation-toolkit repo.Additional Notes
Final toolkit self-review:
approved=False,todos_remaining=18— mostly low-priority comment/naming-hygiene findings and one spurious version-alias flag (the coupled family'ssuper("kafka", "kafka-0.11")alias is no longer justified oncekafka-clients-3.8doesn't exist as a sibling on this branch), not architectural defects.Related:
Jira ticket: [none]