Build an alert for when nothing happens at all

The scariest kind of automation failure is silence. If a process is supposed to run every day and it just stops running, nothing errors out — there is simply no output, and that is easy to miss.

This matters because most alerts only fire when something goes wrong inside the process. They cannot catch the case where the process never starts at all, because a trigger broke or a schedule got disabled. A separate check — something that expects to hear from the automation regularly and speaks up if it goes quiet — catches the failures that regular error alerts cannot see. I've had an automation go quiet on me before I noticed, and the silence itself was the only clue anything was wrong.

You can build this with almost anything: a simple check that looks for a recent log entry and flags you if too much time has passed since the last one.

Try this today: pick your one automation that runs on a schedule and set up a separate check that flags you if it goes quiet longer than expected.