Repository navigation
ci: preserve owned routing during Blacksmith outage - #15614
Conversation
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: Repository: manaflow-ai/cmux/.coderabbit.yaml Review profile: ASSERTIVE Plan: Advanced Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
559cfe6
into
ci/pickers-account-capacity
What changed
When
CI_CLOUD_OVERFLOW_SAVEDrecords an active Blacksmith outage, the PR picker now:MACOS_RUNNER_PRvalue;retry_runnerempty so retries stay on the owned route.The owned-pool rescue watcher pauses while the record exists, avoiding a cancellation that would re-run work onto the dead Blacksmith pool. Normal picker and rescue behavior resumes when the record is removed and the runner variables are restored.
Validation
git diff --checkpasses.This is a dependent PR for #15609 and intentionally contains only the outage integration.
Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is disabled.Summary by cubic
Keeps the PR picker and owned-pool rescue on owned runners during a Blacksmith outage so work doesn't re-run onto a dead pool.
CI_CLOUD_OVERFLOW_SAVEDrecords an active Blacksmith outage, the picker accepts the switchedMACOS_RUNNER_PRvalue, preserves owned split, root, GUI, and warm placement, removes Blacksmith labels from the candidate order, and leavesretry_runnerempty.Validation
git diff --checkpasses.Written for commit 38fbea5. Summary will update on new commits.