Restart a scenario execution that failed with the same input

[sebastiendesautel]

Hello,
I can't find any information how I can re-run/restart a scenario execution that has failed. I sometimes have an error with a scenario due to an impossible connection (Excel module, Excel unavailable). I don't have any error handling. 
I would like to re-run the scenario with the input of the first module which is a webhook. The only solution so far is to relaunch the payload via Postman, but maybe I'm missing a simpler solution.
Is it possible at all?

Thank you
Sébastien

[Lawson02 · Accepted solution]

I have the same scenario with an excel module. I use a break module with incomplete executions setting set to on. This works for me.

[sebastiendesautel]

Thanks for this idea @Lawson02 ,
I will have to learn about error handling.

In the meantime I could find an interesting setting in the scenario settings that stores the incomplete executions in a separate table and they can be completed later. I made a test and I was able to complete a task that failed.

But anyway, is there anything else to restart the scenario in case I have neither an error handling nor the above mentioned setting activated? 

Thanks

Sébastien

[DimitrisGoudis]

Hey @sebastiendesautel, 

You can do it only if your trigger is a webhook. To do this following the steps below: 

- Open form the history the ran your want to copy

- Open trigger's results, download the output, and copy the entire result

- Open Post man or Reqbin, these ar ethe services you can use to trigger the webhook with the same data

- If you use reqbin click on the body tab and paste the trigger's output. Be sure you do not include the first and last symbols: [ ]

- Go back to your scenario on editing mode, opent your webhook trigger, and copy the URL address

- Paste this address into the Reqbin top URL

- Run your scenario once and go back into Reqbin and click the button send. 

Please let me know if that was helpful

[Sven-iX]

@sebastiendesautel 

By turning on "store incomplete executions" you make it possible for @Lawson02's approach to work. Fusion will automatically retry N times with Y minutes between tries. For transient errors such as you list, this is the correct approach. I always try to avoid incomplete executions - and I've had cases where the execution can't be salvaged (e.g. due to unexpected incoming data) and have it turned on only for the BREAK module usage.

If a simple BREAK module doesn't work (or you want to avoid marking the execution as "incomplete") you can roll your own retry and then park the original request bundle in a data store for later retry (since your scenario starts with a webhook).

I've used a flavor where I create my own "retry" logic so that I can, on the last failed retry, create a notification or take another action. The approach involves a repeater and branching success/error paths. Not necessary unless you want to be able to have a "finally" option to your error handling.