Understanding Kubernetes Architecture and Its Core Components

Kubernetes has become the dominant container orchestration engine, offering features like cluster scaling, rolling updates, self‑healing, and service discovery.

The overall architecture consists of a Master (including API Server, Scheduler, Controller) that interacts with etcd for state storage, and Worker Nodes that run containers, kubelet, and kube‑proxy.

When a Deployment is created, the process flows through kubectl → API Server → etcd, then the Deployment controller creates a ReplicaSet, which in turn creates Pods. The Scheduler binds unassigned Pods to suitable Nodes, and kubelet on each Node manages the Pod lifecycle while kube‑proxy sets up service load‑balancing rules.

Pods are the smallest deployable units, grouping tightly coupled containers that share network and storage via an infra container and shared volumes.

Services provide stable endpoints for Pods, using label selectors (e.g., app=xxx) to build endpoint lists and enable load balancing. Service types include ClusterIP (VIP via iptables/ipvs), NodePort, LoadBalancer, and Ingress for external traffic.

Networking in Kubernetes relies on three basic connectivity guarantees: Node ↔ Pod, Pod ↔ Pod on the same Node, and Pod ↔ Pod across Nodes, typically achieved with CNI bridges like flannel.

Overall, the article walks through Kubernetes concepts from architecture to deployment, pod management, service discovery, and external access, equipping readers to understand what is being discussed when Kubernetes is mentioned.