diditarrive
Sign inSign up

Use case

The job you do control, and the night it does not run at all.

A nightly pipeline, a backup, a CI job. It works, so nobody watches it, and the only signal it ever produces is the one it fails to send.

How it fails today

A scheduler is disabled during a deploy, or a machine reboots and the timer does not come back. There is no error, because nothing ran to produce one.

Discovered by whoever needs the output, which is the worst possible reader of that news.

What we do about it

A night, as we would run it
  1. setupOne line in the script: a request to a URL we give you when it finishes.
  2. 02:00Due passes with no call.
  3. 02:30Grace ends. One alert, saying the job never checked in.
  4. 02:58The call arrives after a manual re-run. One closing alert.

What we need

One line in the job. No credential of yours reaches us, because there is nothing for us to log in to. The full posture.

one requestno credentialnothing to open up
The monitor, configured

Nightly pipeline check-in

Where
a URL we generate for this monitor
When
daily by 02:00 UTC
Grace
30 minutes

If the key expires, that’s Check broken, never Missed - broken access is its own alert, and it says whose job it is to fix.

Set it up in a sentence:

“Our nightly job should ping when it finishes, every day by 2am UTC, 30 minute grace.”

Create monitor
Same shape, different plumbing:The connectorsThe schedule modelPricing
Menu
ProductUse CasesConnectorsPricingTrustSign inSign up