TopolearnApps go down and data is sometimes bad; a reliable Zap plans for failure instead of failing silently.
Steps fail for ordinary reasons: the other app is down for a minute, a required field was empty, a connection expired. Zapier gives you three tools.
Replay. From Zap history you can replay an errored run, either only the failed steps or the entire run with the original trigger data. A full replay counts as a new run and uses tasks. Fix the cause first, otherwise the replay fails the same way. Big changes to the Zap's structure can prevent replaying older runs.
Autoreplay retries failed steps automatically a few times, which handles short outages. It is an account setting, and on some plans it can be overridden per Zap.
Custom error handling. On any action step you can add an error handler: a separate path that runs only when that step fails. The original steps sit under "Success" and the handler under "Error". In the handler you might post the app's error message (it can be mapped) to a Slack channel, log the item in a table for review, or try another app. A handled run shows as Stopped/Handled rather than Errored. Notes:
- Not on the trigger or on a Paths step itself (but on steps inside paths).
- One handler per step; handlers can contain paths and more handlers.
- Autoreplay is turned off for Zaps with handlers, and Zapier doesn't email you when a handler runs, so make the handler notify someone.
Also know: by default Zapier turns off a Zap that errors on almost every run over several days, and it emails you about errors. Read those emails.
Remember
- Replay re-runs failed runs after you fix the cause
- Autoreplay retries short outages automatically
- Error handlers give a failing step its own recovery path
Watch out for
- Adding an error handler that swallows failures quietly, so nobody ever learns that orders were missed.