How to check whether an automation worked

The hardest thing about a background job is telling "nothing to report" apart from "not working". Every card answers that in one line, and the run history behind it answers the rest.

What does the health line mean?

The line under each card's title, summarising the last run:

• Never run. It has not run yet. If you only just switched it on, press Run now.
• Last run 12 Aug, 3 found, 3 notified. It worked and told three people.
• Last run 12 Aug, no alert. It ran and there was nothing worth saying. This is the healthy state most of the time.
• Last run 12 Aug, failed with the reason. Something went wrong and the reason is named.

No alert is the one people misread. A low-stock watchdog reporting no alert every fifteen minutes is working perfectly: nothing is low.

What is in the run history?

The ten most recent runs on that card, each recorded as ok, error or skipped, with the counts.

Found, notified and emailed are three separate numbers, and the gaps between them are informative.

• Found 4, notified 0 usually means the recipient list is empty or nobody holds the required access. See How to choose who gets an alert and how often.
• Notified 3, emailed 0 with Also email on means the email did not send. The in-app alert still arrived, and an email failure never blocks it.

What does a skipped run mean?

That the job could not do anything useful, and said so rather than pretending to run.

Two reasons are common, and the run itself names which. A module is switched off, so its automations are dormant. Or the data source is not connected, meaning Shopify or Klaviyo.

A third is the recipient one above: nobody could receive the alert. In that case the alert re-fires later rather than being dropped.

Skipped is not failed, and it is not lost work.

How do you test one?

Press Run now and watch the health line change.

That is the whole loop, and it takes seconds. Switch an automation on, run it now, read the line, and you know whether it works rather than finding out in a week.

Run now goes through the same queue as the schedule, so it cannot double-fire. It does trigger the whole tick, so other due jobs may run at the same moment and their lines may update too.

Why has an alert stopped arriving?

Work down in this order, because each step is cheaper than the next.

1. Is the amber banner up? Automations are off for the whole environment and nothing runs.
2. Is the card switched on?
3. What does the health line say? A failure names its reason, and so does a skipped run.
4. Are you on the recipient list, and do you still have store access? Losing access silently stops the alert even with your name ticked.
5. Is it arriving in your Inbox but not on your phone? Then the automation is working and the problem is a personal notification setting. See What are notifications?.

That last split is the one worth internalising. The Inbox tells you whether the automation worked; the phone tells you whether your own settings let it through.

Why an automation or workflow did not run walks the same path with the workflow cases added.

Why is the run history short?

It holds the ten most recent runs, which is enough to see a pattern and short enough to read.

For a job that runs every few minutes, that is the last half hour. Judge those jobs by the health line rather than by the table.