Sometimes, I get an error on a ChatGPT timeout. (Or other timeouts on other APIs) In this case, I ignore and the scenario stops.
But how can I set up a loop that will allow me to restart the module in the event of an error?
you can insert a sleep module to wait a few seconds (3-5s)
then make a clone of the original module
then use a Resume module and map the cloned module’s outputs to it
You can clone the module that could fail, and add it to its own error handler to “retry” immediately or after a short delay/sleep (in the same execution), then use a Resume module to continue where it failed.