Description:
When a triggered workflow fails, I don't know a direct way to retry the workflow with the original input parameters after fixing the underlying issue. Users must manually re-enter the example parameters from the failed run, which becomes time-consuming when dealing with multiple failed events.
Feature Request:
Add a "Retry" button or option for failed workflow runs that would:
Preserve the original input parameters from the failed run
Allow execution of the fixed workflow using these preserved parameters
Maintain the audit trail between the original failed run and the retry
Use Case:
When a workflow fails due to a code issue
After fixing the code, users need to process the original events
Multiple failed events need to be reprocessed efficiently
Benefits:
Saves time in recovery scenarios
Reduces risk of data loss from triggered events
Improves workflow debugging experience
Maintains data integrity through the retry process
Would this be something the team could consider adding to the platform?
I agree it is tedious to re-enter input data for a workflow when a request fails.
I can definitely make a feature request for this and keep you updated on the ticket's status!
I really like this idea and think it would be very helpful.
The tricky part is that workflows can be triggered by any request being made to them, the workflow would likely need to store the trigger.data payload in a DB, then if the workflow fails several steps later, it would need to update the DB so the payload would have a corresponding 'status' of failed. Which could be done as a work around currently, but would be much better if we have this functionality built in natively. The DB size would be a limiting factor.
I am trying to think about how users would be able to pick which of several trigger payloads with 'failed' statuses that they want to re-run. Maybe some type of queue in chronological order of when the failed attempt was logged
It would be be straightforward to have a 'retry last failed trigger payload' button. But if there are a large number of requests coming in and failing on a live deployed workflow, we would have to build out some type of GUI to view all the failed saves in the 'temp DB'
Let me know your thoughts on these and I can add them to the ticket!
There is a hacky workaround for this. You can go to the failed run > startTrigger > data and then scroll to the bottom, you will find a "Use as example JSON".
This will basically put the data from that workflow in your initial startTrigger block. After that, you can manually play un run workflow so that this is logged.
With regards to the other use cases you mentioned, I agree with you. In fact, I submitted a similar feature request here.