I have seen some posts, but there is no other way to look better than replicating through kafka.
But I do not want to introduce Kafka for this problem.
The second solution is "cross-cluster replication"
I am not sure whether there is a problem in use.
Is there any other solution?