Monitoring for your n8n instance.
n8n runs your most important automations, but it won't tell you when one starts failing, gets deactivated, or the whole instance goes quiet. Keel watches from the outside so you're never the last to know.
n8n doesn't watch itself
There's no built-in dashboard that says "this workflow has been failing since Tuesday" or "your instance stopped responding an hour ago." Without external monitoring, problems surface downstream — usually when a client or a customer notices first.
What Keel monitors
Failing executions
Surfaces workflows that start erroring, with counts, so a quietly-broken automation doesn't run red for days.
Offline detection
A lightweight agent heartbeats continuously; if it goes silent, Keel flags the connection offline and alerts you.
Change & deactivation
Edited, deactivated, or unexpectedly published workflows are caught and diffed in plain English.
Outbound-only, secrets stay home
The agent runs next to your n8n, holds the API key locally, and forwards only metadata — no inbound ports, no payloads.
Find out what changed before your client does.
Keel runs a lightweight agent next to your n8n — snapshots, diffs, alerts, and rollback, with your secrets and API key never leaving your box. Free forever on one instance.
Questions
Can I monitor a self-hosted n8n instance?
Yes — Keel's agent is built for self-hosted n8n. It runs next to your instance and only makes outbound HTTPS calls, so it works behind a firewall with no inbound ports.
How fast will I know something broke?
The agent polls on a short interval and alerts are delivered out-of-band, so you typically hear within a minute or two of a change or failure.
Does monitoring slow down my n8n?
No — the agent reads metadata via the n8n API on a schedule; it doesn't sit in your execution path.
See every change before it breaks a client.
Point Keel at your n8n estate this week. Free 14-day trial, no card, your keys never leave your infrastructure.