Repository navigation
ci(cmux-next): the placement summary names why the DEV compile slot took nothing - #25936
Merged
Merged
Conversation
…ook nothing fresh_slot_state() lists the slot's busy and offline runners (or that none carries the label, or that runners went unread), so a busy slot is told from one its glaeda gate holds offline without the org runners API. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Agent-Session: https://cmux.com/deeplink/agent?id=aid-cc8bff1c3ddb4ec353c5
teamleaderleo
temporarily deployed
to
ci-routing-agent
October 27, 2026 05:15 — with
GitHub Actions
Inactive
Bugbot needs on-demand usage enabledBugbot uses usage-based billing for this team and requires on-demand usage to be enabled. A team admin can enable on-demand usage in the Cursor dashboard. |
Collaborator
Author
|
Owned by an agent session · session |
Contributor
|
cmux status (one comment, edited in place; each line is updated by the job that owns it) Review App of this head (app.zip, kept 1 day)
|
teamleaderleo
temporarily deployed
to
ci-routing-agent
October 27, 2026 05:15 — with
GitHub Actions
Inactive
|
Found 1 test failure on Blacksmith runners: Failure
|
Collaborator
Author
|
Bead |
Contributor
|
Shipped in:
|
This branch was previously 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
![Fix with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/fix-with-codesmith-light-3x.png)
The macOS placement summary now says why the DEV compile slot (
FRESH_LABEL,glaeda-dev-std-xcode-26.6) took no job. It lists the slot's busy runners and its offline runners. If no runner carries the label, or the runners could not be read, it says that instead.Why
Every DEV daily dispatched since #25903 had its
cmux-scheme-compilerefused at "Set up runner" on the shared side label. Both placements, run 40449388179 at 04:23Z and run 40450702256 at 04:37Z, logged only "no idleglaeda-dev-std-xcode-26.6runner". From that line you can't tell whether the slot was busy or held offline. With this change the line shows which.At those times both slot runners were offline, and their glaeda gates were holding them off:
cmux8s-mac-mini-glaeda-2: "only 1 of 14 capacity units on this mini are free; its next job needs 4".cmux11s-mac-mini-glaeda-2: "disk: 21.6 GiB free, 32 GiB required", and then "a fleet build holds the host lock".At both placements Blacksmith macOS also counted as stalled: the oldest queued job had waited about 5 hours. So the compile stayed on the side label, as designed.
The slot did take one compile, run 40452488452's at 04:56:57Z. It queued behind a push's Release compile until 05:06Z. Then the slot's own hook refused it right away: "cmux8s-mac-mini-glaeda-4 release-compile … has waited longer here for units".
What this does not change
Routing is unchanged. The slot is still taken only while idle. Queuing a DEV compile on a busy slot would have no fallback, because the rescue does not watch
glaeda-dev-*labels (pr_runner_pool.OWNED_LABELdoes not match them). A slot can also go offline right after its job finishes, as cmux8s did at 05:07:50Z. The offline gates and the hook refusal belong to glaeda.Risk
Low. Only the summary text changes. The
fresh_runnerandfresh_jobsoutputs, and every other output, are the same.tests/test_ci_side_lane_placement.pyadds a case for the new message.Agent session: agent session
🤖 Generated with Claude Code
Need help on this PR? Tag
@codesmith-botwith what you need.