***Cluster IP - Inner Workings
Hi guys, could someone explain me how does a Pod to Service(Cluster IP) communication works?
Per my understanding so far the Service doesn’t exists as a load balancer somewhere, let’s say in the control panel. If that’s the case e.g. how does a Pod 1 from Front End communicate with Pod 10 from Back End with Service Cluster IP 10.1.1.1 port 80 let’s say?
Does the kube-proxy on each node keep a list of all Service’s Endpoints and round robins outgoing Pod’s 1 requests through them? If so how does Pod 3 from same Front End but different node than Pod 1 keeps a track/sync of all this round robin? Or it just doesn’t matter and each node’s kube-proxy is completely independent?