There are scenarios where many clients send watch requests to the etcd server. When network partition or abnormal restart occurs on the etcd nodes, the leader will be lost on thenodes, and there will be no watch stream connections. Even if the node are healthy, the watch streams among the various nodes will become highly unbalanced. Is there a mechanism that can balance the watch streams of each etcd node?
What would you like to be added?
There are scenarios where many clients send watch requests to the etcd server. When network partition or abnormal restart occurs on the etcd nodes, the leader will be lost on thenodes, and there will be no watch stream connections. Even if the node are healthy, the watch streams among the various nodes will become highly unbalanced. Is there a mechanism that can balance the watch streams of each etcd node?
Why is this needed?
balance the watch streams