galera applier threads not respect innodb_lock_wait_timeout

Even with low innodb_lock_wait_timeout values, the lock waiting time is still 50 seconds.

If in rare corner cases a deadlock happened on replica nodes , which not happened on the write node, master node got "InnoDB: WSREP: BF lock wait long for trx"

after 50 seconds,

------- TRX HAS BEEN WAITING 50000703 us FOR THIS LOCK TO BE GRANTED:

even if innodb_lock_wait_timeout is lower.

This is counter intuitive and the wait lock time for applier threads should be somehow configurable.

I created as a bug report, but it might be converted into a feature request , if needed.