Q: How do I know my monitoring service itself didn't quietly die?

Q: How do I know my monitoring service itself didn't quietly die?

A: Great instinct, because a silent monitor is worse than none, it gives false confidence. The pattern that catches this is a heartbeat (also called dead-man's-switch) check: your system actively pings the monitor on a schedule, and the monitor alerts you when the expected ping doesn't arrive.

That inverts the logic. Normal checks alert when something fails; a heartbeat alerts when nothing happens, which is exactly the failure mode you can't otherwise see, including the monitoring pipeline itself going down.

Use it for cron jobs, backups, and queue workers too: if the nightly backup doesn't check in by 4am, you get paged. No news is bad news.

Follow-up: for the monitoring vendor specifically, subscribe to their own status page and keep one cheap second-source check on a critical URL. Two independent eyes beat one.

Got a question? Drop it in the comments.