Skip to content

test(extension): scrape until the job metrics settle - #35

Merged
juicycleff merged 1 commit into
mainfrom
fix/telemetry-scrape-race
Oct 18, 2026
Merged

juicycleff merged 1 commit into
mainfrom
fix/telemetry-scrape-race

Conversation

@juicycleff

Copy link
Copy Markdown
Contributor

The v1.9.0 release run failed in its verify step on one flaky test, TestExtensionRecordsThroughForge. The tag got pushed, but there's no GitHub release, so this goes out as v1.9.1.

The test scraped /_/metrics once, the moment it saw the process job span end. The worker sets dispatch_jobs_active back to 0 just after that span ends, so a fast scrape could still read 1. Nothing is wrong with the metric itself.

The test now scrapes until every series it expects is there, for up to 5 seconds, and then checks. It passed 200 runs under -race, and another 100 with GOMAXPROCS=2 on a loaded machine. No other test scrapes once like this.

After merge:

gh workflow run release.yml --ref main -f tag=v1.9.1

The v1.9.0 release run failed on TestExtensionRecordsThroughForge. The
test scraped once, right after it saw the "process job" span end, but the
worker drops dispatch_jobs_active back to 0 just after that span ends, so
a quick scrape could still read 1. The test now scrapes until every
expected series is there (up to 5s) and then checks. It passed 200 runs
under -race, and 100 more with GOMAXPROCS=2 on a loaded machine.
@juicycleff
juicycleff merged commit 1fc3f40 into main Oct 18, 2026
17 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant