n8n error alerts
n8n ships an answer for this: the error workflow, a workflow that starts on failure and sends a message. It works — but it has three catches once you look after several client instances.
It has to be set up everywhere
Build it once per instance, then reference it per workflow in the settings. A new workflow without that entry never reports anything. Duskwatch covers an instance completely the moment it is connected.
It only sees what runs
The error trigger fires when a workflow runs and fails. If the schedule stops, nothing runs, and nothing fires. That exact case is the one that costs you clients.
It dies with the instance
If n8n is gone, the error workflow is gone too. A watchdog on the outside notices the outage on its next pass and reports it.
How Duskwatch reports
By email to the agency, additionally to Slack or Telegram. Within a configurable quiet period Duskwatch collects every incident on an instance into a single message — with workflow, node, truncated error message and a link straight to the execution in n8n. Anything that resolves itself before the message goes out is never reported at all.