N8N SILENT-WORKFLOW WATCHDOG CHECKLIST
Agent Rescue Desk
https://agent-rescue-desk.netlify.app/n8n-workflow-stopped-silently-dead-man-switch

PURPOSE
Detect a workflow that never starts, stops polling, or misses its required
business outcome without producing a normal failed execution.

Do not include passwords, API keys, OAuth tokens, private URLs, raw customer
payloads, authorization headers, or unrelated personal data in this worksheet.

1. WATCHED OUTCOME

Workflow name:
Workflow ID:
Environment:
Trigger type:
Expected cadence:
Workflow timezone:
Last verified business outcome at UTC:
Observable success evidence:
Maximum normal duration:
Grace period:
Calculated overdue threshold:

2. THREE DISTINCT CONTROLS

[ ] Error workflow is configured for automatic execution failures.
[ ] Outcome reconciliation compares expected inputs with completed outcomes.
[ ] A dead-man switch expects a heartbeat by a documented deadline.
[ ] Instance-level liveness is monitored outside the n8n failure domain.

3. HEARTBEAT CONTRACT

Stable workflow key:
Heartbeat emitted only after:
Stored fields:
  - workflow key
  - environment
  - last verified success timestamp
  - expected cadence
  - grace period
  - non-sensitive correlation ID
State location:
External ping destination:
Recovery signal:

[ ] Heartbeat data contains no credentials or source payload.
[ ] Test and production use different heartbeat identities.
[ ] The heartbeat proves the business outcome, not merely trigger arrival.

4. OVERDUE RULE

overdue_at =
  last_verified_success
  + expected_cadence
  + maximum_normal_duration
  + grace_period

Business-hours or calendar exception:
First alert threshold:
Escalation threshold:
Repeat suppression rule:
Incident close rule:

5. FAILURE-DOMAIN CHECK

[ ] An in-instance watchdog is not treated as host-outage detection.
[ ] The external monitor can still run if n8n, its scheduler, database, or
    network path is unavailable.
[ ] The alert destination has been tested independently of n8n.
[ ] A unique external ping URL is treated as a bearer credential.

6. CONTROLLED FAILURE DRILL

Test workflow or approved maintenance window:
Expected missed run at UTC:
Expected overdue alert at UTC:
Actual alert at UTC:
Alert count:
Recovery workflow published at UTC:
New verified outcome at UTC:
Heartbeat advanced at UTC:
Recovery signal at UTC:

PASS

[ ] Exactly one overdue incident appeared at the expected threshold.
[ ] No irreversible work was replayed automatically.
[ ] A new verified business outcome advanced the heartbeat.
[ ] Exactly one recovery signal appeared.
[ ] Error-workflow behavior was tested separately with an automatic failure.

HANDOFF

Remaining blind spot:
Rollback:
Owner:
Next drill date:

Independent guide. No n8n affiliation implied.