Skip to main content
A heartbeat rule waits for a job to check in after each run; it fires when no check-in arrives within the period plus the grace, or when a check-in reports a failure.

Check in

The address is given the first time a heartbeat rule is saved; see Create a rule. Request it when the job succeeds; add ?status=fail when it fails, and the rule fires at once instead of waiting for the period. The start of a POST body (up to 1000 characters) is kept as the check-in’s message and shows in the incident annotation heartbeat_message; a body over 64 KiB answers 413. A check-in judges a round of the rule at once: a failure fires it, and a success lets an open incident recover under the recovery condition.
Only status=fail counts as a failure; any other value, or none, counts as a success. A wrong token, a deleted rule, or a reset address answers 404.

Read the check-in state

A rule that is not a heartbeat rule, or has no address yet, answers 404.

Reset the check-in address

Issues a new address; the old one stops working at once, and the check-in record is kept. Returns { "path": "...", "url": "..." }, the address given only this once, with url null when no external address is set. The reset is recorded in the audit log. A rule that is not a heartbeat rule answers 400.
Last modified on October 31, 2026