Issue Summary
This is reproducible on Data Center: (yes)
Steps to Reproduce
#None available, it happens randomly in the backend.
Expected Results
PlanExec, scheduler and ActiveMQ threads are not blocked
Actual Results
Change detection and message handling slows down due to threads being blocked on Java lock.
Taking a thread dump can show Java deadlock involving ExpirableResettableLazyCloseableReference
Workaround
Restart of the application clears the deadlock.