n8n reliability guide
n8n workflow audit checklist
A production-focused checklist for failure paths, retries, duplicate actions, credentials, webhooks, observability and recovery.
1. Map failure paths. Define behavior for timeout, 4xx/5xx, malformed input and partial success.
2. Classify retries. Separate retryable failures from permanent failures before enabling automatic retry.
3. Verify idempotency. Replayed webhooks and reruns must not duplicate messages, tickets, CRM writes, payments or database mutations.
4. Preserve execution evidence. Keep timestamps, correlation IDs and structured errors sufficient to reconstruct incidents.
5. Audit credentials. Keep secrets out of workflow JSON and test fixtures; use credential references.
6. Review risky nodes and entry points. Pay extra attention to code execution, filesystem access, community nodes and public webhooks.
7. Test stale and malformed data. Nulls, missing fields, old timestamps and schema drift should fail predictably.
8. Test duplicate triggers. Webhook retry, polling overlap, queue replay and manual reruns are common duplicate paths.
9. Separate development and production. Test workflows should not point at production credentials or customer destinations.
10. Run a recovery drill. Break one dependency, restore it and verify recovery without double-processing.
What the audit should produce
A useful audit ends with a bounded defect list, reproduction evidence, severity, retry-safety guidance, concrete fixes and a regression plan.
Fast preflight
Export workflow JSON, inspect connections and webhook paths, review credential references, identify state-changing nodes and define how failures are surfaced. For a real incident, pair the workflow export with failed-execution data.
Need a second pair of eyes?
Use the free deterministic preflight first, or buy a bounded review for one workflow.
1-workflow audit — $49Toolkit — $29Free preflightRedacted or synthetic evidence only; do not submit credentials, tokens or customer records.