Repository navigation
fix(scale-down): don't let busy runners use idle slots - #5566
Open
sarankumar1325 wants to merge 1 commit into
Open
sarankumar1325 wants to merge 1 commit into
sarankumar1325 wants to merge 1 commit into
Conversation
A busy runner was counted against idleCount before the busy check in removeRunner, so it consumed an idle slot without being kept as idle. That left the warm pool short and terminated an idle runner further down the list. Only count idle runners against idleCount. Fixes github-aws-runners#5564
This branch has not been 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
In
evaluateAndRemoveRunners, a runner that has passed its minimum running time is counted againstidleCountbefore we know whether it is busy. A busy runner is never removed (removeRunnerchecks the busy state again before deregistering), so it ends up taking an idle slot without being kept as an idle runner. The warm pool is then one short, and an idle runner further down the list gets terminated.This change only lets an idle runner fill an idle slot. The busy state comes from the GitHub runner list that is already fetched for that instance, so it adds no API calls. Busy runners still go through
removeRunner, which keeps them as before.Fixes #5564
Test Plan
Should not let a busy runner use an idle slotto theWhen idle config definedsuite (runs for both eviction strategies). WithidleCount: 1, one busy runner and one idle runner, the idle runner is kept and nothing is terminated. The test fails onmainand passes with this change.yarn vitest runinlambdas/functions/control-plane: 15 files, 501 tests pass.eslint,prettier --checkandtsc --noEmitare clean on the changed files.Note: #5536 also touches
scale-down.ts, but it deals with re-reading the busy state before deregistering, which is a separate problem. If one is merged first, the other may need a small rebase.Related Issues
I used an AI assistant to help write this change. I reviewed and tested it myself.