Break Testing is more than just “force fail.” It is creating tests that will fail for various reasons such as latency, invalid responses, incomplete/unacceptable data… the list of what break testing should test is long.
The surest way to find out what needs to be “break tested” is by first looking at the code and finding out what is “invalid”. Second, look at the requirements documentation and see what is “unacceptable.” If you test for proper handling of both invalid and unacceptable conditions, you will find your code much better accepted in UAT.
As an aside (but an important one) be sure you have documented the type of responses that will occur in invalid and unacceptable circumstances. Make sure that documentation is in layman’s terms (simple, non-technical) and make sure the user community understands and approves of how the exceptions are handled.