Keep them; an n8n Error Trigger catches an unhandled failed run, and you wire it per workflow, per client, per platform. Nothing inside the platform catches a run that never happens, and when the instance goes down, the error workflow goes down with it. Everrun watches the platform's own record from outside, across every client, without adding anything to the workflows or scenarios you built.

Read, and nothing else: every call to n8n or Make is a GET, with no code path that writes to, edits or triggers an automation, so a key scoped to read is all it needs. Credentials are encrypted at rest with AES-256-GCM and decrypted at poll time, and the Security page lists the controls in full.

We never fetch or store execution payloads — only status metadata and sanitised error messages, with credentials stripped. One honest qualifier: error text is authored by your own workflows, so it may incidentally contain personal data.

The client page: every automation you run for them, its state, its last run and, on Growth and Scale, the hours it saved this month. The alert email is written to be forwarded; it names your agency, the automation and what the platform recorded, in plain words.

Add one Webhooks by Zapier step at the end of each Zap to send a JSON payload to the URL Everrun gives you; it's a Premium step, so the Zap needs a paid plan. Zapier has no public run-history API, so that ping is the signal, and a Zap that stops pinging inside its expected window is the alert.

Email privacy@everrun.dev for an export or a deletion; both are written procedures we run by hand, answered within a month. Deletion removes the account and its clients' data, encrypted backups are kept for up to 90 days, and a downgrade deletes nothing.