Errors (or faults) can occur when executing a process. Using fault handlers you can catch faults or exceptions and create fault-handling procedures to deal with potential runtime errors in your process definitions.
Fault handlers are defined at the scope level, as a result you can catch faults or exceptions thrown by activities within a scope. To catch faults or exceptions specific to an individual activity, you need to define a new scope for that individual activity and attach a fault handler to the new scope.
At runtime, once a fault handler is executed, the associated scope will not complete due to the error thrown. If a fault is not thrown in the fault handler, the process execution continues with the first activity that follows the scope. If a fault is thrown in the fault handler, then the engine looks for an enclosing scope that is designed to handle the fault. If one is found, the engine executes it. Once the enclosing fault handler finishes its execution, the engine executes the next activity following the scope. If no fault handlers are found in the enclosing scopes, then the job terminates with a fault.