The engine is not working as expected in a event subprocess when an event subprocess throws a end throw event such as a Escalation Throw End Event or Error Throw End Event. This seems to effect all end events in the event subprocess.
Seems this entire pattern is unusable i.e. thorwing events in event subprocess.
The example below illustrates the issue. Notice two tokens exist in the event sub-process because the timer has fired twice even though the end throw event has been reached on the first try and should have canceled the event sub-process. Also notice that the parent token has been canceled on the first cycle while the catch event was correlated on the first cycle as well.
The engine does not end the event subprocess leaving a token at the End Event, but a new token continues into the catch event leaving a hanging token in the process.
The token in the end event of the subprocess should be terminated, terminating the subprocess.