Kubernetes · Orchestration under pressure
How is a Kubernetes cluster architected?
Short answer
A cluster has a control plane that stores and exposes state, and worker nodes that run pods. The API server is the hub, etcd is the source of truth, and per-node agents execute work.
The control plane runs the API server, etcd, the scheduler and controller managers. Worker nodes run kubelet, a container runtime and a networking component. Every component talks to the API server rather than to each other directly, which keeps the system loosely coupled.
- kube-apiserver validates requests and is the only component that writes to etcd
- etcd stores the entire cluster state and needs consistent backups
- kube-scheduler assigns unscheduled pods to nodes
- controller-manager runs built-in reconciliation loops
- kubelet starts containers and reports pod status on each node
kubectl get pods -n kube-system -o wide
kubectl get nodes -o wide
kubectl cluster-infoManaged offerings hide the control plane but not its behaviour. You still need to understand API server availability, etcd backups, admission control and version skew, because those determine what happens during upgrades and incidents.
Follow-ups you should expect
Interviewers use these to check whether the first answer was understanding or memorisation.
- 1Why does every component talk to the API server?
- 2What happens when the API server is unavailable?
- 3How does a pod get from the API to a running container?
- 4What is the difference between the control plane and data plane?
Say it in the interview
A cluster splits into the control plane and worker nodes. The API server is the front door, etcd persists state, the scheduler picks nodes and controllers reconcile objects, while kubelet and the runtime run pods on each node. Components communicate through the API server, which keeps them decoupled. That design also means etcd durability and API server availability are the things I watch most closely.
Use this as a model, not a script. Saying it in your own words is what makes it survive follow-up pressure.
Related questions
Reading is step one
Now say it out loud.
Understanding an answer and delivering it under pressure are different skills. Rehearse this topic in a live Kubernetes mock interview and get scored feedback on depth, clarity and confidence.