How to configure the keep-alive parameters for the store backend #924
Description
How to configure the keep-alive parameters for the store backend? For example, if I configure --store-backend as kubernetes, I want the instance pods (such as proxy, keeper, etc.) to quickly release the connection and reconnect to other available nodes when a Kubernetes node becomes unavailable. Currently, I notice that when I power off a Kubernetes node, it takes the proxy pod 15 minutes to reconnect to a functioning Kubernetes node.