Well, that is a tough question, because it highly depends on what exactly triggers the block.
I assume there is some sort of an interface, where you can look up past blocks by their ray IDs. Finding the block by that ray ID should clarify what exactly triggered the firewall to block the whole request. Then it should offer a way to make an exception for that request type.
Usually it’s something to do with the request body – something in the POST variable, maybe an argument it resembles a threat in the heuristics database.