I’ve built a small monitoring service for n8n workflows and wanted to share it here.
What it does: each workflow sends a one-line HTTP ping when it finishes. Kept works out the schedule from those pings, asks you to confirm it, and then alerts you (email, Slack, Discord or Telegram) if a run is late, doesn’t happen, or fails. Alerts include a link to the execution.
The gap it covers: the Error Trigger fires when a workflow fails, but not when a workflow doesn’t run at all. Kept notices the missing run.
For failures, there’s an error-reporter workflow you import and set as your Error workflow. It sends six fields to Kept: workflow name and id, execution id and URL, the failing node’s name, and the error message. No item data.
Pings are metadata only, limited to 2 KB, and email addresses, phone numbers and tokens are scrubbed before storage.
It’s my own project. Free for three workflows, £9 a month for 25. Site: trykept.uk
Happy to answer questions about how it works.