Currently the testing will fail approx once in 10k runs. It has been traced back to a timeout issue as the timeout is set as 1ms but the timeslice in linux is 10ms so it's possible for some code to get stuck while waiting for another slice. The length of this timeout is non critical as it's only used in shutting down threads, so it can safely be extended.
Success criteria
Timeout extended to avoid threading issues
Timeout set in single place rather than once per class.
Currently the testing will fail approx once in 10k runs. It has been traced back to a timeout issue as the timeout is set as 1ms but the timeslice in linux is 10ms so it's possible for some code to get stuck while waiting for another slice. The length of this timeout is non critical as it's only used in shutting down threads, so it can safely be extended.
Success criteria