Skip to content

ci(runners): send twelve default-runner jobs to the kirocrew-runners group - #25645

Merged
bolichen97 merged 1 commit into
mainfrom
ci/scheduled-lanes-runner-group
Oct 31, 2026
Merged

bolichen97 merged 1 commit into
mainfrom
ci/scheduled-lanes-runner-group

Conversation

@tim-finnigan

@tim-finnigan tim-finnigan commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Move twelve jobs across ten workflows off GitHub's shared standard
runner pool onto the kirocrew-runners larger-runner group.

Why it matters

The standard GitHub-hosted pool shares its concurrency enterprise-wide, so a
job on it waits behind work this repository does not control. The owner's call
is that these lanes should not sit there.

Not a goal

  • Making anything faster. The group is the same CPU and memory as a
    public-repository ubuntu-latest.
  • Changing any check name, trigger, permission or step.
  • Moving build-desktop.yml's build-desktop job or
    build-ffmpeg-decoder.yml's build job. Both pin ubuntu-22.04 and
    ubuntu-22.04-arm to match the fleet image's glibc 2.35 floor; that is a
    compatibility decision of its own.
  • Moving any job's macOS or Windows leg. The group is Linux-only.

What changed (motivation → approach → change)

Twelve jobs asked for a shared runner; now they ask for the group this
repository pays for. Nothing else about them changes.

Workflow Job(s)
main-red-watch.yml, main-red-pr-notice.yml, main-macos-failure-watch.yml watch, notice, page
issue-summary.yml, issue-triage.yml summarize, triage
nightly.yml version, process-leak-invariant, pod-scenarios (Linux leg only)
pr-readiness-sweep.yml sweep
gui-user-test.yml gui-user-test
build-desktop.yml smoke-linux-packages
build-ffmpeg-decoder.yml publish

Eleven were a hardcoded runs-on: ubuntu-latest and become:

runs-on: ${{ github.repository == 'kirodotdev/KiroCrew' && 'kirocrew-linux-4-core' || 'ubuntu-latest' }}

The label, not group:. A group resolves only inside the organization that
owns it, so a fork running its own copy of one of these workflows finds no
runner and the job waits forever with nothing to fall back to. The label form
carries its fallback in the || leg. That costs the rename safety a group:
reference buys: rename the definition and these twelve lines break together.
The larger-runner totals in docs/ci/ci-and-reviews.md are left to
PR #26303, which removes
those running totals, so this PR edits no doc.

nightly.yml's pod-scenarios is a three-leg matrix. Its name: and its
os: values are untouched, so the rendered check names do not move; only the
Linux leg's runner is decided, per leg, the way ci.yml's e2e-boot-matrix
does it.

Nothing gets faster, and that is the honest framing: same CPU and memory as
a public-repository ubuntu-latest, so only queue position changes, and
larger-runner minutes are billed where standard ones are free. These lanes hold
no merge, so this is not a latency win being bought.

One exception is worth a reviewer's attention. pr-readiness-sweep is how a
completed lane reaches the PR Readiness verdict at all, because
pr-readiness.yml does not subscribe to the completed event of the workflows
it monitors. A sweep that waits for a runner delays a green verdict reaching
the required status. It cannot make a verdict wrong, and cannot turn a red into
a merge, but for that one job queue position is the latency.

Backwards compatibility

Nothing is rejected, renamed or removed. Check names, triggers, permissions and
steps are identical, and a fork keeps ubuntu-latest through the || leg.
Rollback is a one-line revert per site.

Tests

test/test_ci_fleet_routing_expression_parity.py pins every off-fleet job to
the exact runner it is expected to use. Three of these jobs are already
_PERMANENT_EXCEPTIONS entries, so their _HOSTED_EXCEPTION_RUNNERS pin moves
from the ubuntu-latest default to the new expression. The pin stays an exact
per-job equality, so a drift to a third runner still fails.

test/test_main_macos_failure_watch.py asserts its watcher's runner by
equality; that assertion now names the new expression. It is re-pinned, not
relaxed, for the same reason: a drift to a self-hosted runner, where "no
checkout" stops being the whole isolation story, must still fail.

No new test. The property both files already pin is exactly the one this diff
changes, and a-ratchet-may-only-tighten wants it re-pinned rather than
widened.

Manual verification

python3 scripts/local-gate.py backend leg on the rebased head: 1265 passed,
118 skipped across the 14 related files. Its frontend leg could not run on this
host, which failed npm ci with a registry E401 before any check; no
node_modules exists on this machine for any checkout, and this diff contains
no website/ file.

Both changed test files re-run alone and serially,
python -m pytest -n0 test/test_ci_fleet_routing_expression_parity.py test/test_main_macos_failure_watch.py: 34 passed, 26 skipped. The
flake-pattern ratchet, which the related-test runner does not select, is also
green: python -m pytest -n0 test/test_flake_pattern_ratchet.py, 64 passed.

Every one of the twelve runs-on values was re-read back out of the parsed
YAML and compared against the intended expression, together with
pod-scenarios's unchanged name: and os: list and the two deliberately
untouched matrix jobs.

Checked at the changed/unchanged boundary: scripts/ci/runner_watchdog.py acts
only on labels carrying its codebuild- prefix, so it neither heals nor
notices a kirocrew-linux-4-core job, and no job here gained a fleet label.
runner.os conditions inside pod-scenarios still read Linux on the moved
leg. No moved job declares a container: or services: block.

Screenshots / video

Why no screenshot: CI runner routing only; no frontend file is touched and
no pixel changes.

Related Issues

Closes #25644

Checklist

  • Local gate run; backend green, frontend leg blocked by a host
    npm ci failure that reproduces on every checkout on this machine
  • Check names unchanged
  • No doc edit here; the larger-runner totals in
    docs/ci/ci-and-reviews.md are PR #26303's to remove

@github-actions

github-actions Bot commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision b4b539113a722a5dba92110a2d99d6a8366eff7e touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 29, 2026
@github-actions

github-actions Bot commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of b4b539113a722a5dba92110a2d99d6a8366eff7e — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks. It reuses the fork-safe label form already used in 31 other workflows, the check names stay the same, and each site can be reverted with a one-line change.

Notes

  • The same kirocrew-linux-4-core ternary is now written out at about 46 sites. If the runner definition is renamed, every one of them breaks at once, and the parity test will flag it.

[DESIGN-REVIEWED] b4b5391

Description read

[DESCRIPTION-READ] f5a1a10252b0cd261e779070455a2b5efc012d38c6add4e85a57bd933d4d3a3a

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of b4b539113a722a5dba92110a2d99d6a8366eff7e — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

This only moves CI runners, and the PR is explicit that a fork falls back to ubuntu-latest. The label expression it uses is already in 31 workflows. No item changes product shape.

Notes

  • Item 13 — duplicate of an inline literal in the same dict: the new _GROUP_LABEL_EXPR sits next to 9 entries that spell the same ${{ github.repository == 'kirodotdev/KiroCrew' && 'kirocrew-linux-4-core' || 'ubuntu-latest' }} string inline (test/test_ci_fleet_routing_expression_parity.py:206-230). That leaves two spellings of one pin in one table. Subtraction: delete _GROUP_LABEL_EXPR and write the 3 new entries inline like their neighbours.

What this change ships

Inventory (13 items) — 12 justified

Intent: keep this repository's maintenance and nightly CI lanes out of the shared, enterprise-wide standard runner queue. ADDITION (an operational change).

  1. The main-red-watch watch job runs on the kirocrew-runners group in the upstream repo — justified
  2. The main-red-pr-notice notice job moves to the group — justified
  3. The main-macos-failure-watch page job moves to the group — justified
  4. The issue-summary summarize job moves to the group — justified
  5. The issue-triage triage job moves to the group — justified
  6. The nightly version job moves to the group — justified
  7. Only the Linux leg of nightly pod-scenarios moves; check names stay the same — justified
  8. The nightly process-leak-invariant job moves to the group — justified
  9. The pr-readiness-sweep sweep job moves to the group — justified
  10. The gui-user-test job moves to the group — justified
  11. The build-desktop smoke-linux-packages job moves to the group — justified
  12. The build-ffmpeg-decoder publish job moves to the group — justified
  13. The test pins are re-pinned to the new expression, adding a _GROUP_LABEL_EXPR constant — duplicate of the inline literals in the same dict

[FIRST-PRINCIPLES-REVIEWED] b4b5391

Description read

[DESCRIPTION-READ] a75b3a917b2af037406d5da079d3c151b4bf7dcb7458337408e511cfe9283d7a

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of b4b539113a722a5dba92110a2d99d6a8366eff7e and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] b4b5391

False positive or not applicable? A repository writer can comment:
/ai-review override gpt b4b539113a722a5dba92110a2d99d6a8366eff7e false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Oct 29, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed b4b539113a722a5dba92110a2d99d6a8366eff7e — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] b4b5391

Verdict parsed from the review's SHA-scoped output markers for commit b4b539113a722a5dba92110a2d99d6a8366eff7e.

False positive or not applicable? A repository writer can comment:
/ai-review override fable b4b539113a722a5dba92110a2d99d6a8366eff7e false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: checking Automated validation is still running labels Oct 29, 2026
@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

@tim-finnigan this branch now has merge conflicts with main. Please rebase it onto main and push. The lanes on the current head all pass, so it can be approved once the conflict is gone.

@tim-finnigan
tim-finnigan force-pushed the ci/scheduled-lanes-runner-group branch from 56f0d51 to 3a85dbf Compare October 29, 2026 06:15
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 29, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor

HOLD at 3a85dbf86720c6c70234d6087920c5f3577004dc.

The change itself looks sound: twelve runs-on values move to the repository-conditional kirocrew-linux-4-core label with an ubuntu-latest fork fallback, pod-scenarios moves only its Linux leg, and no permission, secret or trigger changes. The two runner pins in tests are re-pinned by equality, not relaxed.

Blocker: Build ffmpeg decoder is red on this head. build (macos-aarch64) failed with curl: (28) Failed to connect to ffmpeg.org port 443 and the other three legs were cancelled. It looks like a network flake (the same workflow passed on the previous head), but it is the only pull_request run of the workflow this diff edits, and it is not one of the lanes PR Readiness monitors, so auto-merge would not wait for it. A rerun of the failed jobs that comes back green clears this.

Non-blocking AI notes, for the record: the Opus note that the doc says build-desktop.yml declares only workflow_call when it also has workflow_dispatch is a wording slip whose conclusion still holds; the Design label-vs-group rename note and the First Principles doc-length note are documented tradeoffs/size nits.

— crew-pr-reviewer (r135)

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 29, 2026
@tim-finnigan
tim-finnigan force-pushed the ci/scheduled-lanes-runner-group branch from 3a85dbf to 3f9e9ea Compare October 29, 2026 18:12
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 29, 2026
@tim-finnigan
tim-finnigan force-pushed the ci/scheduled-lanes-runner-group branch from 3f9e9ea to 3c83cbb Compare October 29, 2026 21:26
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: action required A blocking check or review needs attention and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 30, 2026
@tim-finnigan

tim-finnigan commented Oct 30, 2026 •

Copy link
Copy Markdown
Contributor Author

self-added: no
mechanism: routing-only diff; the doc inventory edit is tracked in #26425
resolution: deferred to #26425

  • Doc line 1116 still names these six workflows as staying on ubuntu-latest (span=7818d10f0bb8)

accepted-and-deferred to #26425: the finding is correct and is not being
rebutted. docs/ci/ci-and-reviews.md:1090 does require the owning entry and
subsection to be edited by a routing change.
It is out of scope HERE because all five open routing PRs must append to the
same inventories (the group: list near 1075-1089 and the label-route lists
near 1363-1403). Carrying the edit per PR puts the five back in conflict with
each other and serializes the wave behind repeated rebases, which is the exact
state these PRs were just cleaned up to escape.
#26425 makes that edit once for every landed wave. It carries the
deferred-finding label, an assignee and a Due date, and it names this PR's
sites and the false sentences explicitly.

@tim-finnigan

Copy link
Copy Markdown
Contributor Author

/ai-review override fable 29831b2 deferred #26425: agent: owner-directed deferral. The finding is correct and is not rebutted. All five open routing PRs append to the same ci-and-reviews inventories, so a per-PR doc edit puts them back in conflict with each other. #26425 makes that edit once per landed wave.

@github-actions

Copy link
Copy Markdown
Contributor

Human judgment recorded

@tim-finnigan recorded a deferred release (class deferred) for the fable AI finding on 29831b24b83faadeada8a2f69222fbdcaad8ac2e.

agent: owner-directed deferral. The finding is correct and is not rebutted. All five open routing PRs append to the same ci-and-reviews inventories, so a per-PR doc edit puts them back in conflict with each other. #26425 makes that edit once per landed wave.

This decision applies only to this commit. A new push requires a new judgment.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: action required A blocking check or review needs attention labels Oct 30, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor

Hold: one product question about pr-readiness-sweep and the runner cap.

Reviewed head: 29831b2

The change itself is sound. Twelve runs-on sites take the fork-safe kirocrew-linux-4-core label expression already used on main, no permission, secret, trigger or step changes, and the two runner assertions are re-pinned by equality. The doc inventory gap the Opus lane found is tracked in #26425, which names this PR's sentence. CI is green on this head, including the Build ffmpeg decoder run that was red at r135.

The open question: your sibling #25349 measured kirocrew-runners at peak 102 / p95 98 against maximum_runners: 100 and asks that merge-path lanes not land until the cap is raised. It also leaves pr-readiness-sweep off the group on purpose. This PR moves that same job (

runs-on: ${{ github.repository == 'kirodotdev/KiroCrew' && 'kirocrew-linux-4-core' || 'ubuntu-latest' }}
), and its description says queue position is exactly the job's latency to PR Readiness. Please confirm whether the cap blocker applies here. Either answer clears this: the cap has moved, a maintainer accepts landing it now, or the sweep is dropped from this PR.

AI findings judged non-blocking: Design (label literal repeated across workflows, rename cost) does not block because the parity test pins every site, so a missed rename fails CI.

— crew-pr-reviewer (r137)

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Oct 30, 2026
…group

Twelve jobs across ten workflows move off GitHub's shared standard pool onto
the `kirocrew-runners` larger-runner group, named by its runner label with a
fork fallback rather than by `group:`, because an organization group is
unreachable from a fork and a job targeting one there waits forever for a
runner.

Nothing gets faster: the group is the same CPU and memory as a
public-repository `ubuntu-latest`, so only queue position changes, and
larger-runner minutes are billed where standard ones are free.

Check names are unchanged, including nightly's `pod-scenarios`, whose `name:`
and `os:` values stay exactly as they were with the decision moved into
`runs-on` per leg.
@tim-finnigan
tim-finnigan force-pushed the ci/scheduled-lanes-runner-group branch from 29831b2 to b4b5391 Compare October 31, 2026 00:50
@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Oct 31, 2026
@bolichen97
bolichen97 enabled auto-merge (squash) October 31, 2026 02:22
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.

ci: route twelve default-runner jobs to the kirocrew-runners group

4 participants