Repository navigation
alerting: retry pending Full CI analyses hourly between scheduled ticks - #136
Merged
Merged
Conversation
A failed Full CI analysis previously waited for the next twice-daily (05:00/19:00 PT) tick, so one failure could delay a report by up to 12 hours and go unnoticed. Add an hourly retry timer that drains pending comparisons between scheduled ticks: - alerting-full-ci-retry.timer fires at *:43 UTC; the service runs only 'run-worker full-ci-analyze' (reconciliation stays on the Pacific schedule) and skips while the scheduled run is active so retries never duplicate an in-progress analysis. Concurrent duplicates are also harmless: commit_analysis upserts with ON CONFLICT DO NOTHING. - The control service enables/disables the retry units with the rest of the full_ci path, so a 'disabled' mode stops retries too. Signed-off-by: Thang Nguyen <thang.nguyen@inferact.ai>
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
This branch was successfully deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Problem
A failed Full CI analysis only retried at the next twice-daily tick (05:00 / 19:00 PT), so one failure could delay a report by up to 12 hours and go unnoticed — this happened today with the #87024 nightly report.
Change
alerting-full-ci-retry.timerfires hourly at*:43UTC and runsrun-worker full-ci-analyzeonly — reconciliation stays on the Pacific schedule. The analyzer already drains pending comparisons oldest-first and is idempotent (commit_analysisupserts withON CONFLICT DO NOTHING), so more frequent invocation is safe.alerting-full-ci.serviceis active (ExecStartPreguard) so a retry never duplicates an in-progress analysis.full_cipath, so setting the mode todisabledstops retries too.alerting/docs/full-ci.md).Tests
full_cimode (enabled when live/shadow, disabled+stopped when disabled).pytest alerting199 passed,deploy/aws24 passed, ruff clean, mypy shows only the pre-existing local-venv boto3/stub errors (identical without this change).Deploy
Takes effect on the worker after the next
install.shrun frommain(ops handbook §9); the control service enables the new timer within a minute of the unit files landing.Signed-off-by: Thang Nguyen thang.nguyen@inferact.ai