Synthetic Checks for n8n Webhook and Worker SLAs
Uptime checks are not enough for n8n. A load balancer can return healthy while workflows are stuck, Redis is degraded, or workers cannot call a required API. Synthetic checks give you a way to test the full automation path on a schedule.
For security workflows, this matters because missed automation can become missed response.
Context
Problem: Basic HTTP health checks do not prove that webhook intake, queue processing, and downstream actions work. Approach: Run safe synthetic workflows and measure each step as a service-level signal. Outcome: Reliability issues show up before a real alert depends on the platform.
Synthetic workflow design
A good synthetic workflow is safe, cheap, and representative:
- Trigger through the public webhook path.
- Validate request authentication.
- Enter the same queue path as production workflows.
- Execute on a worker.
- Write to a test ticket, test queue, or dedicated audit table.
- Emit timing metrics.
- Clean up after itself.
It should never touch production accounts, disable users, or mutate real incidents.
Metrics to track
Measure:
- Webhook response time.
- Time from webhook accepted to job active.
- Worker execution duration.
- Downstream test action duration.
- Success rate.
- Error category.
- Release ID active during the check.
Use separate thresholds for business hours and off-hours if your worker scaling changes by schedule.
ECS integration
Synthetic failures should be correlated with:
- Recent ECS deployments.
- Worker desired and running count.
- Task stopped reasons.
- Redis and database errors.
- ALB target health.
- CloudWatch Container Insights metrics.
This makes the alert actionable. “Synthetic failed” is only useful when the first responder can see the likely layer.
Blue team use
Security teams can use synthetic checks as tamper and drift signals. A workflow that suddenly stops emitting expected test events may be broken, disabled, blocked by network policy, or changed without review. Pair checks with change logs to narrow the cause.
Takeaways
Synthetic checks turn n8n reliability into an end-to-end measurement. Test the path a real alert takes: webhook, queue, worker, downstream action, and audit output.