DB failure during TX in WIH spawns error handling process

A customer provided a reproducer (simulator) for an issue they claim to see in production. The high level overview is that a process with a custom WIH (async) runs into a OptimisticLockException, as a result the engine executes the error handling sub-process, which also contains an async node. In this scenario, the thread executing the parent issues a lock on the child process, which is never released, see details attached (full log files and thread dump in the archive). The thread executing the async node in the child process is then blocked waiting for the lock.
Engineering's current thinking is a database problem should not trigger a process handling error as it leaves the workflow in an inconsistent state. They've opened bug report https://issues.redhat.com/browse/JBPM-10011 failure database during tx in WIH causes spawn a process error handling.