Managing IP addresses of Kubernetes worker nodes

Could you please explain how to manage changes in the IP addresses of Kubernetes worker nodes? The monitoring stack, connectors, and other clients rely on these IPs, but they need to be updated every time the node pool undergoes maintenance.

With hostNetworking and ephemeral IP addresses, you need a script when your client starts that scrapes those from the kubernetes API. I’d recommend not using hostNetworking mode though.

With Service IPs, or Pod IPs this becomes more managable to scrape, but even better is to use the IP for service <scyllacluster_name>-client that proxies to all nodes, and the client will discover the other nodes on its own.

Thanks. How is this service configured and can it be exposed as load balancer?

I found the documentation about exposing ScyllaCluster, but the client service is not mentioned.