Currently we have the 'wait-store-timeout' configuration item to protect the DR state in 'sync' from short-term jitters.
As we know, either the DR state is 'sync' or 'async', the DR replicas are in ACID consistent state. Unless the DR state is changed to 'sync_recover', we don't guarantee the DR replicas' ACID consistency.
So, we need a timeout in 'async' state to protect the DR replicas' ACID consistency from short-term jitters.