Overview
An automation is only useful if you can trust it. Run history shows what each run did, and retries absorb temporary failures such as a mail server that is briefly unavailable.
Who can do this
- Required access: an access level that includes Workflows - Standard Access or Workflows - Full Access.
- Default access levels that include it: System Administrator.
- Owners and admins: always allowed. Workspace owners and admins are not limited by access levels.
- Neither the Team Member nor the Accountant access level includes Workflows, so by default only owners and admins see the module. Add a Workflows level to an access level to open it to others.
- Personal accounts: Not available. This is a work feature, so it is hidden in personal (family) spaces.
Owners and admins review and change access levels in Settings → Administration → Members → Access levels.
Step 1: Review Run History
- Open the workflow in Workflows and click Run history.
- Review the latest 20 runs. Each shows a status (Pending, Running, Completed, Failed, or Skipped) and per-step logs.
- Open a failed run to see which step failed and why.
Step 2: Add retries to fragile steps
- Click the action node that failed.
- Turn on Retry on failure.
- Set Max attempts (2 to 10, default 3) and Wait between (minutes) (default 5).
- Click Save.
A failed run can't be re-run from the app. Fix the cause, then let the next trigger run it, or start it with a test.
Step 3: Pause, rename, or delete
- Pause (in the header) stops an Active workflow and keeps its setup.
- In the workflow list, use a row's … menu to Rename it, or Delete it. Deleting also removes its runs and history.
- Use the status filter (All statuses, Draft, Active, Paused, Archived) to find workflows.
Troubleshooting
| Issue | What to check |
|---|---|
| Runs or steps show Skipped. | Read the log: a path is skipped when its conditions don't match, and test runs skip delayed or scheduled actions. |
| A step keeps failing after retries. | Read the step log: missing required fields and deleted records don't fix themselves with retries. |