Looping back to failed script task calling integration in process models

We have a script task that calls a rule within which we make a few calls to some integrations and then process the returned data and save into a PV and then the flow continue further.
now the problem is the API catering us with the data is not a stable one and fails sometimes causing the script task to fail. I want to implement a loop around in the process model where I should be able to loop back to the script task and retry executing. how do we approach this problem. I want to reduce human intervention here as much as possible.