19-11-22 02:23 PM
21-11-22 12:09 AM
21-11-22 01:22 PM
Eric is correct, three attempts then mark as exception and move onto the next case. Usually a seperate process will then deal with exceptions (either report them out or send them to a team to be worked manually).
I would recommend looking at the process templates available in the portal (developer jumpstart area) if you have not already. The templates include an advance version with more complex exception handling.
Ideally you would only mark items as exceptions if there is a problem with an individual case you want to work (i.e. it is a scenario the robot has not been trained to work), but you would not go through marking lots of items as exceptions if a system is down/unavailable. The advanced template contains logic related to a system being unavailable.
Also, your retry exception handling would ideally include system restart and navigation back to the right location logic - at least on the last of your three attempts.