Github user tillrohrmann commented on the issue:
https://github.com/apache/flink/pull/5744
Thanks for the contribution @yanghua. I fear that the `TASK_KILLING` states
have been left out deliberately. When a task is being killed, then you don't
stop. Thus, the issue is actually a non issue. What we could do is to make the
match exhaustive to add a `TASK_KILLING => noOp`.---
