When a galera node is also an async replication replica, and it encounters a replication error (e.g. a duplicate key), the Last_SQL_Error in "show slave status" output is populated with "Node has dropped from cluster", which is misleading.

Node has not dropped from cluster. It's the cluster that has dropped from upstream replication.

I believe that this error should be removed completely and the actual replication error should be reflected in the Last_SQL_Error field of "show slave status", in other words the behaviour should not be changed regardless if the replica is a galera node or not.

Thanks,
Rick