~/ nmadiraju.com Amazon EKS · Services and kube-proxy
arrow_back Back to nmadiraju
Visual Walkthrough · Kubernetes networking

How Services and kube-proxy Work on Amazon EKS

Follow web's calls to shop through a ClusterIP, kube-proxy and EndpointSlices, then check yourself with 8 questions.

1Visual Walkthrough

A Service, its EndpointSlice and kube-proxy on two EKS nodes, with traffic traced to ready Pods Two nodes, each with kube-proxy. node-1 runs the web caller and shop-a; node-2 runs shop-b and shop-c. The API holds Service shop (ClusterIP 172.20.14.7) and its EndpointSlice. The walkthrough shows a stale Pod IP failing, the ClusterIP and EndpointSlice, kube-proxy rewriting traffic on the sending node, per-connection balancing, a Pod dropping out on failed readiness, no endpoints, and port vs targetPort. node-1 · EC2 · 10.0.1.10 kube-proxy iptables rules for every Service web pod (caller) 10.0.1.22 calls shop on :80 shop-a 10.0.1.15:8080 REPLACED node-2 · EC2 · 10.0.2.10 kube-proxy iptables rules for every Service shop-b 10.0.2.31:8080 REPLACED shop-c 10.0.2.44:8080 REPLACED KUBERNETES API Service shop ClusterIP 172.20.14.7 selector app=shop port 80 → targetPort 8080 shop.default.svc.cluster.local EndpointSlice shop-x7k kept up to date by a controller

    request packet control-plane update address rewritten here Pod or agent object in the API

    Press Play to watch all nine steps, or step through with Back and Next. Red step tabs are where something breaks; green ones add the fix. Voice reads each step aloud with your browser's built-in speech.

    2Quiz to check understanding

    Check yourself