Repository navigation
Re-enable duplicate metric test with default timeout - #6088
Merged
max-charlamb merged 4 commits intoOct 8, 2026
Merged
Conversation
TestMetricsLogger previously wrote 'Found expected counter' and 'Received additional counter event' lines directly through the raw ITestOutputHelper, with no timestamp. This made it impossible to tell exactly when counter payloads arrived relative to the pipeline stop sequence, which is useful for diagnosing intermittent timeouts in TestDuplicateNameMetrics/TestCounterEventPipeline (e.g. dotnet#4757). Add elapsed-time and payload Timestamp/Interval info to each log line so CI runs give exact arrival times for counter events during the stop drain, instead of only being bounded by neighboring timestamped lines.
Increase the process watchdog to six seconds to investigate whether timing explains the Windows and macOS failures tracked by dotnet#5375. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 90d32376-768b-4f8c-a982-54247c372cf3
This was referenced Oct 7, 2026
Rely on the shared 60-second process watchdog rather than a test-specific timeout. Keep Windows and macOS enabled. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 90d32376-768b-4f8c-a982-54247c372cf3
max-charlamb
changed the title
Keep the PR focused on removing the platform skip and using the default process timeout. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 90d32376-768b-4f8c-a982-54247c372cf3
max-charlamb
marked this pull request as ready for review
October 7, 2026 21:00
Member
Author
|
/ba-g legacy test failure |
rcj1
approved these changes
Oct 7, 2026
Member
Author
|
/ba-g legacy test failure |
This was referenced Oct 12, 2026
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
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.
Summary
Fixes #5375.
Remove the test-specific three-second process watchdog from
TestDuplicateNameMetricsand use the shared 60-second default, matching the other pipeline tests. Re-enable the test on Windows and macOS. The .NET 8 skip remains because instrument IDs require .NET 9 or later.Investigation
The original logs did not timestamp counter callbacks, which made the delay look like slow shutdown. Instrumented runs show that most time is spent waiting for counter delivery, not stopping the pipeline.
With a temporary six-second watchdog, all 21 .NET 9/10/11 cases passed across Linux x64/ARM64, Linux musl x64, Windows x64/x86/ARM64, and macOS ARM64 in build 1626709. The macOS ARM64/.NET 11 run received its last expected counter at 3.347 seconds and finished at 3.597 seconds, exceeding the original three-second watchdog. macOS x64 did not run this test project.
Using the standard watchdog leaves room for collection intervals and CI scheduling delays without delaying successful tests: the pipeline stops as soon as all four expected tag combinations are received.
Validation