Due to the repeated intermittent failures, and the subsequent discussion on the problem being attributed to lazy subsystem activation. Instead of disabling them all (per policy), I have merged an update which isolates the OSGi tests to their own configuration with eager activation. The drawback to this is that OSGi tests are not included in "smoke" tests. However, we always run the full test suite on every commit and PR anyway. We can now monitor the situation and see if intermittent failures occur again.