The Error Workflow catches runs that fail. It can’t catch a workflow that stops running, because there’s no execution to fail. This workflow covers both from one place, without editing the workflows you watch.
Every 15 minutes it reads your instance’s executions through the n8n API and sends one Slack message (or Discord, via the /slack webhook) when:
- any workflow has a failed or crashed run since the last check
- a workflow you list hasn’t run for longer than you allow, e.g. a daily job with no run in 25 hours
Setup is in the sticky note: an API key as Header Auth, then your instance URL, a webhook URL and the list of workflows to watch in the Config node. It only reports a quiet workflow once, then again after it recovers and goes quiet again.
Limits, so nobody’s surprised:
- It reads the last 250 executions. On a busy instance a quiet workflow can fall outside that, and it skips it rather than guess.
- Keep “Save successful executions” on for watched workflows, or it can’t see their runs.
- It runs inside the instance it watches, so it can’t tell you n8n itself is down. Pair it with an outside uptime check.
- It doesn’t catch runs that go green with 0 items. For that, put an IF (no items) into Stop and Error right after the step that must return data. The watchdog then reports it as a failure.
Copy this and paste it into a blank n8n canvas:
Happy to fix anything that doesn’t work on your version.
Disclosure: we make RunLookout, a paid monitor for n8n. This workflow is free and doesn’t need it.