Repository navigation
ci: run the required checks for merge-queue entries - #1020
Merged
Merged
Conversation
CI, CodeQL and dependency review run for merge_group, and the title and signature checks report for it, so the main ruleset can require a merge queue. A queue entry gets the complete validation that main gets. Pull requests can then leave the three recovery database groups to the queue. That is controlled by mergeQueueDatabaseGroups in the rollout policy on main, which stays empty until the queue is required.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What changed and why
Pull-request runs wait in line because one CI run starts about 30 jobs and the organization may run 20 at once. The recovery database groups (
crash-points,configuration-recovery,coordinator-recovery) are 12 of those jobs and about 144 of a full run's roughly 250 job-minutes, and they run on every push to every pull request. This prepares a merge queue so they can run once per merge instead. Nothing changes for current pull requests until the ruleset requires the queue.merge_group.Validate conventional commit titleandcheck / record signaturereport without repeating work that needs a pull request, because every queued pull request already passed them.merge_grouplike a push: scopefull, every database group, no pull-request-only Docker build.mergeQueueDatabaseGroupsinscripts/ci-impact-rollout.jsonon main lists them. The list is empty in this change. Only those three groups can be listed, the aggregate gate checks shard receipts against the same plan, and a deferral on any other event is rejected.After this merges: require the merge queue in the
main protectionruleset (settings are in the governance runbook), then a one-line follow-up lists the three groups. Main still runs the full suite after each merge; reusing the queue's result there is a possible later step.Verification
merge_groupevent, produces scopefull, no deferred groups and all 17 database partitions.merge_grouptrigger, job names and unique concurrency group of the five workflows. The seven CI contract and plan test files pass (88 tests), as dopnpm lintandpnpm check:policy.merge_grouprun exists until the ruleset requires the queue, so the five workflows' behaviour on that event is untested on GitHub. The dependency-review action is given the queue's base and head commits explicitly; its README documents those inputs for other events but does not namemerge_group. The runbook says how to back out if a check stays pending.actionlintandzizmorare not installed locally and run in CI.Release and upgrade impact
Product impact: ordinary