I'm seeing too many requests for a write operation on a node when I query it with the thread pool API. On other nodes, this status appears as 0. It seems like write requests come to only one node. How can I fix this situation?
Also when there are too many write requests the nodes can leave the cluster or I get the following error.