Repository navigation
🐛 fix: explain Activity Monitor merge gates and sweep timing - #15081
thatguywob wants to merge 2 commits into
Conversation
Distinguish reviewed, ready, soaking and merging rows, expose merge-sweep clocks and preserve gate context on nested PRs. Label the gate glyph and agent activity separately. Hive-Run: hivecommons#15073 Hive-Plan: merge-gate-observability Hive-Spec: issue-15073#fix Co-authored-by: clubanderson <407614+clubanderson@users.noreply.github.com> Signed-off-by: thatguywob <thatguywob@users.noreply.github.com>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
@thatguywob this PR is back in your court ( Failing checks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
Signed-off-by: thatguywob <thatguywob@users.noreply.github.com>
Open PR overlap checkBase branch:
Conflicts are checked with every PR merged onto the current tip of
|
|
@thatguywob this PR is back in your court ( Review asks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Queued in your relay fix queue (1 waiting): fixes are handed out before new work as soon as you have capacity. |
|
PR needs rebase. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether #15081 or #15084 lands for #15073 — both implement the issue's Fix section and both create src/pkg/github/automerge/sweep_clock.go and src/pkg/dashboard/stage_tracker_merge_gate.go, so they cannot both merge.
This appears to duplicate #15084, which implements the same #15073 Fix section against v5. I read both diffs in full against their v5 merge-bases.
What I verified is the same in both:
/api/stage-trackerPR items gainmerge_gatewithstate,reason,sweep_running,last_sweep_at,next_sweep_at,soak_until,soak_remaining_seconds(here via the embeddedgithub.MergeSweepTimesinstage_tracker_merge_gate.go:11-17; in #15084 as flatomitemptyfields).- Identical
STAGE_TRACKER_ACTIVE_WORDmap instatic/index.html— 060 Awaiting merge gate, 080 Ready for auto merge, 095 Soaking, 100 Merging. - A sweep clock armed by the automerge lanes, feeding the dashboard.
Where they differ, so they conflict rather than merely overlap — both create the same two new files with different contents:
- This PR keeps the clock in
pkg/github(merge_sweep_clock.go, aMergeSweepClockfield ongithub.Client,client.go) and reaches it fromautomergethrough a transport type-assertion (sweep_clock.go:9-15); #15084 keeps aSweepClock/SweepStateinsidepkg/github/automerge(sweep_clock.go) and snapshots targeted PRs per pass. - #15084 also updates
src/docs/api-reference.mdfor theserver.gochange and ships a 261-line merge-gate test table; this PR adds thecmd/hivewiring test andmerge_sweep_clock_test.go.
Both build (go build ./pkg/dashboard/ ./pkg/github/... ./cmd/hive/). I have not formed a correctness judgement on either beyond that, because whichever one is not chosen is moot. #15084 is authored by the repository owner and is not behind v5, while this one is needs-rebase, so #15084 looks like the natural survivor — but that is the owner's call. I have not closed anything.
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
There was a problem hiding this comment.
Decision needed: @clubanderson should decide whether #15081 or #15084 is the fix for #15073 that lands; both implement the same merge_gate payload and 060/080/095/100 vocabulary in the same files and cannot both merge.
This appears to duplicate #15084, which fixes the same issue (#15073) with the same change. I read both diffs in full:
Same semantic change (verified):
- Same vocabulary: both set
STAGE_TRACKER_ACTIVE_WORD060 → Awaiting merge gate, 080 → Ready for auto merge, 095 → Soaking, 100 → Merging insrc/pkg/dashboard/static/index.html. - Same payload: both add
merge_gatetoStageTrackerItemand nestedlinked_prs(src/pkg/dashboard/stage_tracker.go) with the same field names —state(gate-pending|ready|soaking|merging),reason,sweep_running,last_sweep_at,next_sweep_at,soak_until,soak_remaining_seconds(here:stage_tracker_merge_gate.go:11-17+pkg/github/merge_sweep_clock.go:11-15; in #15084:stage_tracker_merge_gate.go:36-52). - Same files for the clock hooks:
automerge_sweep.go,automerge_forgestate.go,sweep_clock.go,dashboard-sections.md, plus a changelog fragment each.
Where they differ — which is why they conflict rather than stack:
- Clock location. This PR adds
pkg/github.MergeSweepClock(per-lane, hung off the client,client.go) and wires it fromautomerge_contributor_lane.goandcmd/hive/main.go; #15084 keeps it insidepkg/github/automergeasSweepClockwith aSweepStatesnapshot that also records the sweep cause and targeted PRs. omitempty: this PR serialises everymerge_gatefield always ("last_sweep_at": nulluntil a pass runs); #15084 omits empty ones.- Docs: #15084 also updates
src/docs/api-reference.mdandlabels-and-control-signals.md; this PR adds acmd/hivewiring test instead. - Merge state: this PR is
dirtyagainstv5; #15084 isblocked(checks), not conflicting.
Both were written from #15073's Fix section, so the overlap is expected, but only one can merge — the second will conflict in all of the shared files above. #15084 is the broader one (sweep cause, API reference) and currently applies cleanly, which points to it as the survivor; the one thing this PR has that #15084 lacks is the automerge_sweep_wiring_test.go check that both the event loop and the ticker arm the clock, which could be carried over.
Flagging for a maintainer: if you agree, this one could be closed in favour of #15084. I have not closed anything.
— hive: agent=reviewer-2 backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Supersession sweep found that this PR's claimed issue is already closed by another merged PR.
Decision needed (@hivecommons): #15084 already merged and closed the claimed issue, but this PR touches files the merged PR did not. Recommended: keep open until the extra changes are reviewed, or close as superseded if they are not wanted. Another option: merge anyway. Files the merged PR did not touch: This comment is edited in place by the supersession sweep; it is not duplicated. |
|
Closing as superseded by #15084 (merged 2026-10-28 20:22Z), which fixed #15073. #15084 covers everything the issue's Fix section asked for: |
A reviewed PR with pending CI currently reads “Ready for auto merge”, and its spinning glyph only means an agent is working. Nested soaking PRs also lose their countdown.
This change gives 060, 080, 095 and 100 distinct words, consistently shows a labelled merge-gate glyph, and labels agent activity separately. Client-owned sweep clocks publish the current PR evaluation, last completed pass and next scheduled floor pass; event-triggered passes can happen sooner. PR items and nested
linked_prscarry merge-gate state, reason and soak timing, and nested soak pills show the time left. Section help, the dashboard guide and a changelog fragment explain the cues. The clocks are observational and do not change merge admission.Validation:
gofmt -lon changed Go files andgit diff --checkpassed. Added regression coverage for all four lanes, concurrent evaluations, queued-pass timing, payload fields, distinct words, accessible glyphs and nested countdowns. Tests were not run locally, as required by AGENTS.md; CI runs them.Fixes #15073
Hive-Run: #15073
Hive-Plan: merge-gate-observability
Hive-Spec: issue-15073#fix (the issue's Fix section)
— hive: backend=codex
🐝 Hive Agent:
contributor| SHA:39a8e50c8