Failing triggers now pause instead of disabling workflows

Poll triggers check a service on a schedule and start a run for each new item. When polling broke, we used to disable the whole workflow after three consecutive failures, whatever the cause, so a provider outage could switch off an automation its other triggers were still feeding. As of April 30, a failing trigger pauses only itself.

How it works

• Retries in place. Network errors and timeouts get up to three attempts within a check, and an authentication error retries with a freshly loaded credential in case the token rotated.
• Judged by status code. Lost access (401, 403) or a deleted resource (404, 410) pauses the trigger after one failed check. Server errors are tolerated for 25 consecutive checks, other client errors for 3. A rate limit never pauses a trigger.
• Only the trigger pauses. The workflow stays enabled, its other triggers keep firing, and you get a notification naming the trigger and the reason. Editing the workflow after fixing the cause turns it back on.
• A shorter lookback. Checks look back one hour instead of 24.

New triggers

Google Sheets row updates, Airtable records, Jira issues, DocuSign envelopes, Ashby candidates, Ramp transactions, Asana tasks, and PandaDoc documents.