Cancel Run
Mark a stuck or running pipeline run as failed. Use when a run has not made progress for more than 15 minutes.
Forces a queued or running pipeline run into the failed state. Use when a run appears stuck — for example, meta.stage_started_at has not advanced for more than 15 minutes — or when you need to free the active-run slot for a new trigger.
The run-status cancellation is the only thing this endpoint guarantees. Background work that was already in flight when you cancelled may continue to completion. The status row is locked into the cancelled state — a worker that finishes after the cancel cannot transition the row back to running or done (PR #112 added a terminal-state guard) — but its outputs (evidence, opportunities, brief drafts, etc.) may still land in the database under the cancelled run_id. Filter by ?run_id= if you need to inspect what the worker produced before it was cancelled.
Path parameters
brand_idstringrequiredThe brand UUID.
run_idstringrequiredThe run UUID.
Response
{
"run_id": "run-uuid",
"status": "failed",
"cancelled": true
}Errors
| Status | Meaning |
|---|---|
400 | Run is not in a cancellable state (already done, failed, or cancelled). Body: {"detail": "Run is <status>, not queued or running"}. |
404 | Run does not exist or does not belong to this brand. |
Example
curl -X POST \
-H "X-API-Key: $BM_API_KEY" \
https://api.boringmarketing.com/brands/$BRAND_ID/runs/$RUN_ID/cancelAfter cancelling, you can immediately re-trigger the same pipeline endpoint. The new run will get a fresh run_id. Pipeline data is append-only — the cancelled run's partial work is preserved in the database and can still be queried via ?run_id=<cancelled-run-id>.