Given the above diagram. How does Load Balancing work in Kubernetes EKS? To my understanding it works like that:
Elastic Load Balancer balances the call between worker nodes (in my case it redirect the call to the Node 1 as shown with the purple arrow)
Then the kube-proxy on the node which was picked by the Load Balancer (Node 1) checks the Ip Tables to route the traffic the correct pod according to a Round Robin mechanism, the pod choosed by the Round Robin can be located also inside another node (since they are all connected to the same overlay network)
QUESTIONS:
Is the flow described above correct?
Is it corrected to say that ELB balances at Node level and Kube-Proxy at Cluster Level?
Which caller ip address sees the final pod? The client ip address? the ELB ip address? or the Node Ip Address where it lies the Kube-Proxy replica that balanced the traffic (so in my case the Node 1 Ip Address)?