Skip to content

Conversation

@delan
Copy link
Member

@delan delan commented Dec 2, 2025

unfortunately neither the new global queueing feature (servo/servo#40852), nor dropping the runner-timeout jobs (servo/servo#40876), have improved self-hosted runner utilisation (a metric of server workload).

that said, the self-hosted runner hit rate is 93%, so not too many jobs are missing out on self-hosted runners.

@delan delan requested review from jschwe and sagudev December 2, 2025 07:19
@delan delan force-pushed the queue-analysis branch 2 times, most recently from 69f3c86 to 7f5fd65 Compare December 2, 2025 08:08
Comment on lines +29 to +30
those workflow runs include 655 job runs that were allowed to reserve self-hosted runners, of which 611 actually used self-hosted runners, which is a 93.2% success rate.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sounds like we should work on allowing more jobs to run self-hosted runners then.

@delan delan merged commit 208b9fc into main Dec 2, 2025
1 check passed
Copy link
Member

@sagudev sagudev left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

and it looks like this got merged while I was reading it

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Well we do profit from having self-hosted runners so that is something.

Maybe we can reduce IDLE time by having running some expensive scheduled jobs (webgpu CTS, wpt runs, coverage). How much idle time is actually "in one slice"?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants