Today, we use a timeout when waiting to ensure a worker thread is queued waiting for a ticket, or is woken up and receives a ticket, in these unittests. These timeout values aren't enforcing any real constraint/correctness property; they're just trying to exit the test early in a case where we would have failed due to a hang where we never waited for a ticket or received one. 

Today we use 100ms as the timeout, which on slow windows variants has caused a couple of timeout failures. We should increase the timeout to some value where hitting the timeout would be unambiguous evidence of a hang and not caused by slow machines to avoid this test noise.