When a step fails, Upstash Workflow schedules the next attempt according to the retry delay of the run. If you've already fixed the issue on your endpoint, you don't need to wait for the next attempt. You can retry the step right away and let the workflow run continue.
Retrying now does not add an extra attempt. It moves the already scheduled retry forward, so the retry count and the remaining retries stay the same. If the step fails again, the next retry is scheduled as usual.
Only steps waiting for a retry can be retried now. If a run failed after all retries, use the DLQ to resume or restart it instead.
Retrying via console#
In your Upstash Workflow console, go to the Logs tab and find the step that is waiting
for a retry. Press the Retry Now button on the right side:
Retrying programmatically#
You can retry a step now using the Upstash Workflow REST API with the message ID of the step.
The request returns 200 when the step is scheduled for immediate execution, and 404
when the step is not found or is not waiting for a retry.
If multiple parallel steps are waiting for a retry, send a request for each of them.
Message ID#
The message ID of a step is available in the run logs. Fetch the logs of your run with
the logs API or
client.logs, and find the step in STEP_RETRY state.
Its messageId is the ID you need.