Run a script only when a custom metric goes down.

Alerts › scripts. Custom metric is the kind for a command that posts ok or down. A failed check is a different kind.

An empty key still means every metric. One metric is a narrower key.

Downtime starts ticked

Open Alerts › scripts and press new trigger. Name it and choose Who is signed in. Downtime is already ticked. Custom metric is clear, so a post that goes down would not start this draft. Disk, checks, reboots, network path and Docker apps are clear too.

New alert script trigger with downtime ticked and custom metric clear — New trigger. Downtime is ticked. Custom metric is clear. Nothing is saved yet.

Tick custom metric and clear downtime

Tick custom metric and clear downtime. The label names a POST that reports ok or down. A missed post, one that never arrives inside the grace, is this same kind. A failing check is not. Accounts server's disk at 92% is a disk alert, so it would not. The key box stays empty, so every metric matches, on every machine. The floor is warn and above, and a down post is a warning unless the metric itself is set to critical. Press save when this is the one you want. This shot stops before that. Cancel drops the draft. Run on is still these machines, and none is ticked, so a save would also need a target.

Alert script trigger with only the custom metric kind ticked — Custom metric is ticked. Downtime is clear. Save has not been pressed.

Telling a person, with no script, is Only tell them about a custom metric. Creating the URL a command posts to is Add a custom metric. A yes or no status is Add a yes or no custom metric. A failed check, which is a different kind, is Run a script only when a check fails.