Sometimes, in order to debug a machine, many tasks may be terminated, and then new tasks will not be received.
If a computer has a very large number of errors, the server will stop sending tasks. However, this requires literally thousands of errors, so unless your debugging went epically wrong that's not what's happening here.
Were you doing your testing on a project with short tasks such as SGS, or one with long tasks such as SoB? If you were using a project like SoB, there's a risk that your testing could have produced an error on each of the workunits that are currently available to be sent. This can happen with fewer than 100 errors. Until someone else grabs some tasks, causing new workunits to be cycled in, you can't be sent any more tasks for those workunits because you're not allowed to have more than one task per workunit and you've already had one (failed) task in all the available workunits.
On a project such as SGS, however, tasks are going out so quickly that there's no problem getting tasks for fresh workunit and you're unlikely to encounter this problem.
My lucky number is 75898524288+1 (I am NOT an administrator anymore, so please don't PM me with questions. I can't help.)