Dead-man switch (heartbeat)

SupaCove integrates with dead-man switches such as healthchecks.io: after every successful AND fresh backup it pings your URL; production-path failures ping <url>/fail.

ConfigurationCopy Anchor Link

In the database's "Schedule & heartbeat" drawer:

• Heartbeat URL: empty inherits the server-wide SB_HEARTBEAT_URL; - disables explicitly (no success and no fail pings).
• Expected period (hours): required when the URL is a real address; with - (disable) no positive period is demanded (while disabled neither success nor fail pings are sent; a stored period value is NOT zeroed automatically and keeps working once you switch back to a real or inherited URL). When inheriting the fallback URL, success pings still require a positive period. Your external monitor alerts when pings stop for period+grace.
• Grace (hours): extra tolerance on the period.

When a success ping firesCopy Anchor Link

1. an effective URL exists (per-database or fallback) and is not -;
2. the expected period is > 0;
3. the time from THIS run's export start (dumpStart) to now is within period + grace — not the completion time, and not a poll of history: a very long successful run may legitimately not ping;
4. with a remote destination configured: that backup must be remote-committed (local-only deployments are exempt).

Fail pings bypass the age gate (the failure IS the signal) but still respect -. Your external monitor's alerting period is configured on the monitor side; SupaCove only pings.