Repository navigation
feat(OMN-19888): scheduled heals update release-window reds and re-run vendor-parity reds once their cause is gone - #3026
Merged
jonahgabriel merged 2 commits intoSep 27, 2026
Conversation
…ty reds once their cause is gone A PR synced inside a release window fails Release Identity Gate and a rerun replays the same merge ref; a PR whose migration was vendored into omnibase_infra dev after its parity gate failed stays red because nothing reruns the gate. pr_stale_red_heal.py decides both from live reads and acts once per head: a branch update pinned to expected_head_sha (App token, so the new head gets CI) when the failed version is below dev and dev is out of the window; a rerun of the parity run plus CI Summary when every named migration is byte-identical on infra dev and was vendored after the gate failed. Onex-Lane: market-heals-83 Onex-Session: 73ed3d50de9e420c9d831fe17a2275a3
…its merge-queue commit date omnibase_infra#4206's squash commit is dated 15:38:41Z and reached dev at 16:01:07Z; a parity gate that failed at 15:53:01Z ran before the bytes landed. Reading the committer date misclassified it as a real red. Onex-Lane: market-heals-83 Onex-Session: 73ed3d50de9e420c9d831fe17a2275a3
|
| Rule | Description |
|---|---|
| ARCH-TOPIC-001 | No hardcoded Kafka topic strings in handler code |
| ARCH-DI-001 | No event_bus=None bypass |
| ARCH-DI-002 | No direct Handler instantiation outside workflow_runner/adapters |
| ARCH-DI-003 | No reinvented DI containers |
| ARCH-TOPIC-002 | contract.yaml topics follow `onex.{cmd |
ARCH-DI-002: Direct Handler instantiation (warning)\n\n```\nsrc/omnimarket/nodes/node_projection_delegation/handlers/handler_delegation.py:1909: row.update(HandlerDelegationCohortKeyFold().handle(event).row_columns())
src/omnimarket/nodes/node_projection_delegation/handlers/handler_delegation.py:1914: ticket = HandlerDelegationTicketFold().handle(event)
src/omnimarket/nodes/node_projection_delegation/handlers/handler_delegation.py:1926: caller_lane = HandlerDelegationCallerLaneFold().handle(event)
src/omnimarket/nodes/node_projection_delegation/handlers/handler_projection_delegation.py:993: row.update(HandlerDelegationCohortKeyFold().handle(event).row_columns())
src/omnimarket/nodes/node_projection_delegation/handlers/handler_projection_delegation.py:1011: ticket = HandlerDelegationTicketFold().handle(event)
src/omnimarket/nodes/node_projection_delegation/handlers/handler_projection_delegation.py:1023: caller_lane = HandlerDelegationCallerLaneFold().handle(event)\n```\n\nHandler classes should be wired only in workflow_runner.py or adapter modules. Direct instantiation elsewhere bypasses the contract system.\n\n
Static architectural lint — no model inference (OMN-14176).
onexbot-occ-writer Bot
added a commit
to OmniNode-ai/onex_change_control
that referenced
this pull request
Sep 27, 2026
…nimarket#3026 The declared checks were executed for real in the OmniNode-ai/omnimarket checkout at the PR head by the product-repo OCC receipt runner, and the results written append-only into this companion. Producer-minted PENDING receipts are superseded by a net-new record; no existing receipt was edited. Ticket: OMN-16859 (AC3b) Onex-Workflow: OCC Receipt Runner Onex-Run: https://github.com/OmniNode-ai/omnimarket/actions/runs/36333868449
jonahgabriel
enabled auto-merge (squash)
September 27, 2026 16:37
✅ Hostile Reviewer — PASSEDBlocking findings (critical/error): 0 Gate semantics
Powered by omniintelligence.review_pairing.cli_review — node-based adversarial review via HandlerLlmCliSubprocess (OMN-8524) |
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
| class GhPort(Protocol): | ||
| """The GitHub reads and writes both heals need.""" | ||
|
|
||
| def open_pull_requests(self, *, repo: str) -> tuple[OpenPr, ...]: ... |
|
|
||
| def open_pull_requests(self, *, repo: str) -> tuple[OpenPr, ...]: ... | ||
|
|
||
| def check_runs(self, *, repo: str, head_sha: str) -> list[dict[str, object]]: ... |
|
|
||
| def check_runs(self, *, repo: str, head_sha: str) -> list[dict[str, object]]: ... | ||
|
|
||
| def job_log(self, *, repo: str, job_id: int) -> str: ... |
|
|
||
| def job_log(self, *, repo: str, job_id: int) -> str: ... | ||
|
|
||
| def job(self, *, repo: str, job_id: int) -> dict[str, object]: ... |
|
|
||
| def job(self, *, repo: str, job_id: int) -> dict[str, object]: ... | ||
|
|
||
| def run_jobs(self, *, repo: str, run_id: int) -> list[dict[str, object]]: ... |
|
|
||
| Any other failure raises :class:`RuntimeError`. | ||
| """ | ||
| ... |
| The merge time of the PR that landed it when there is one, else the | ||
| commit's committer date; ``""`` when unread. | ||
| """ | ||
| ... |
|
|
||
| def update_branch( | ||
| self, *, repo: str, pr_number: int, expected_head_sha: str | ||
| ) -> None: ... |
| self, *, repo: str, pr_number: int, expected_head_sha: str | ||
| ) -> None: ... | ||
|
|
||
| def rerun_run(self, *, repo: str, run_id: int) -> None: ... |
|
|
||
| def rerun_run(self, *, repo: str, run_id: int) -> None: ... | ||
|
|
||
| def rerun_job(self, *, repo: str, job_id: int) -> None: ... |
jonahgabriel
pushed a commit
to OmniNode-ai/onex_change_control
that referenced
this pull request
Sep 27, 2026
* evidence(OMN-19888): author OCC companion for OmniNode-ai/omnimarket#3026 OCC companion by node_pr_lifecycle_fix_effect (OMN-13317 F1 / OMN-13990 / OMN-14285). Product PR head 77c098b612764b78805df10d58bb03952ace5b6e. * evidence(OMN-19888): self-bind OCC#11670 + rebind contract_sha256 * evidence(OMN-16859): executed test_passes receipts for OmniNode-ai/omnimarket#3026 The declared checks were executed for real in the OmniNode-ai/omnimarket checkout at the PR head by the product-repo OCC receipt runner, and the results written append-only into this companion. Producer-minted PENDING receipts are superseded by a net-new record; no existing receipt was edited. Ticket: OMN-16859 (AC3b) Onex-Workflow: OCC Receipt Runner Onex-Run: https://github.com/OmniNode-ai/omnimarket/actions/runs/36333868449 --------- Co-authored-by: omnimarket-bot <bot@omninode.ai> Co-authored-by: onexbot-occ-writer[bot] <307849072+onexbot-occ-writer[bot]@users.noreply.github.com>
jonahgabriel
deleted the
jonah/omn-19888-omnimarket-heals-release-window-vendor-parity
branch
September 27, 2026 17:23
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
2 participants
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.
OMN-19888: scheduled heals for release-window and vendor-parity reds
Ticket: OMN-19888 (child of OMN-19852). Source: the omnimarket stall diagnosis, ledger TERMINAL 2026-09-27T16:07:43Z lane=market-armed-stall-83, findings 3 and 4.
What this fixes
Two required reds on omnimarket PRs are caused by something outside the PR, and neither one clears when that cause goes away.
Release Identity Gate. A rerun replays the same pinned merge ref, so only a branch update clears it (fix(OMN-19234): the bus receipt names the backend key that answered, not uuid5 of the model id #3009, feat(OMN-19513): node_projection_claude_hook_events -- every Claude Code hook, with subagent lineage #2956, feat(OMN-19513): the emit registry routes hook.event and redaction governs nested members by dotted name #2955, feat(OMN-19550): full-content capture topic, a span-scrub capture class and a session_content projection #2905 on 2026-09-27).node-migration-vendor-parity-gatefails until omnibase_infra dev vendors the PR's migration. Nothing reruns it afterwards, andCI Summaryasserts it as an external context.What this adds
scripts/ci/pr_stale_red_heal.pyfollows the shape ofocc_companion_merge_heal.py(OMN-18812). It has two pure decision functions, aGhPortprotocol, and one enumerated outcome per branch. An unreadable open-PR list exits red, and so does a pass where every write fails.update-branchpinned toexpected_head_shawhen all three hold:Release Identity Gatefailed naming a version below dev's current one;CI Summaryjob, when both hold:merged_at, not the squash commit's date. On omnibase_infra#4206 the merge queue dated the commit 15:38:41Z, and it merged at 16:01:07Z..github/workflows/pr-stale-red-heal.ymlruns every 10 minutes, offset 5 minutes from the companion heal. It also runs on push to dev touchingpyproject.toml(the bump PR merging) and onworkflow_dispatch(pr-number, heal, dry-run).pull_requesttrigger and nocontinue-on-error, and it is not a required context.actions: writeis scoped to its one job.release-on-merge.ymluses. An update-branch authored byGITHUB_TOKENstarts no workflow on the new head. Without the token, the update is refused, not issued.tests/ci/test_pr_stale_red_heal_omn19852.pyhas 54 tests, including the ticket's AC falsifiers:-k release_window: 20 pass-k release_window_refuses: 14 pass-k vendor_parity: 22+ pass-k vendor_parity_refuses: 15+ pass-k workflow: 5 passWhy not bulk_pr_throttle.py
Lanes route bulk update-branch and rerun through omnibase_infra's
bulk_pr_throttle.py. This scheduled workflow cannot use it, for two reasons:CROSS_REPO_PATis being retired under OMN-16373.expected_head_sha.The script applies the same bound itself: at most 5 branch updates and 10 rerun PRs per pass, with the rest deferred to the next pass.
Controls (live, read-only; no write was issued)
conflicted.not_vendored. Infra PRs fix(OMN-17427): release staleness counts dev merge tags as published #4210, docs(OMN-17427): remove 33 unreferenced docs files after archiving them #4169, feat(OMN-20910): native Anthropic Messages protocol with prompt caching for the review gateway #4209 and feat(OMN-21144): score recorded gate cascades #4154 are still open.gate_not_red.dev_in_window, so it is not updated while dev is inside the window.update_required, parsed from the live gate log "pyproject version 0.4.248 is NOT ahead of the latest published version 0.4.248".rerun_requiredfor run 36331166338. That is the run a lane reran by hand at 16:15Z after omnibase_infra#4206 merged.failed_after_vendor. That was the defect the second commit fixes.Checks
ruff format/ruff checkclean on the new files.mypy --strictclean on the module and the test.pre-commit run --fileson the three files: rc 0.check_action_pins.pyOK.check_workflow_module_refs.pyPASS.test_ci_summary_gate.py,test_self_hosted_sibling_clone_hygiene.pyand the OMN-18812 heal suite (246 tests).Lab
The change is CI-only: a workflow, a script under
scripts/ci/and a test. It touches nothing undersrc/, docker or migrations. It is not runtime-affecting, so there is no lab pass to cite. There is also no tenant, RLS or reader path.No collision with omnimarket#3023, which edits
occ_companion_merge_heal.py, its workflow and its test. This PR touches none of those files.Evidence-Ticket: OMN-19888
Evidence-Source: OCC#11670