Why did my run fail?
Read the run timeline to see which step failed and why.
Updated
Every run is recorded step by step, so you can see exactly where something went wrong and resume from there.
Open the run
Go to Runs and click the failed run. The timeline shows every step with its input, output, and duration. The failed step is marked in red with its error message.
Common reasons
| Reason | What to do |
|---|---|
| An app was unavailable | Retry; most outages are brief |
| Permissions changed | Reconnect the integration or widen its scope |
| Missing data | Check the record the agent was working on |
| Instructions were unclear | Refine the instructions, then retry |
| An approval was rejected | This isn’t an error; the run stopped as intended |
Retry from the failed step
Click Retry from this step. The run resumes from its last checkpoint, so earlier steps aren’t repeated and no email is sent twice.
Prevent it next time
- Add a fallback to the instructions, such as “if the order can’t be found, hand the ticket to a person”.
- Turn on alerts for failed runs in the agent’s settings.
- Review Analytics → Runs weekly to spot recurring failures.
Tip: if one step fails often, split the agent into two smaller agents connected by a workflow. Smaller agents are easier to debug.
Was this article helpful?
Thanks for the feedback.