Mastering Kubernetes: A Deep Dive into Core Architecture and Components

Kubernetes (K8s) has become the de facto standard for container orchestration. Understanding its core architecture helps design reliable, scalable cloud‑native platforms.

The system follows a classic master‑node (control plane‑data plane) design. The control plane consists of the master node, while the data plane comprises worker nodes.

Master node (control plane)

The master node manages the entire cluster, acting as the brain and command center, coordinating resource allocation and scheduling.

Key components:

kube-apiserver – the unified entry point for all REST operations, handling authentication, authorization, validation, and audit.

etcd – a distributed key‑value store that holds the cluster’s desired state, requiring strong consistency and high availability.

kube-scheduler – assigns unscheduled Pods to suitable nodes based on resource availability.

kube-controller-manager – runs various controllers (ReplicaSet, Deployment, Node, etc.) to reconcile actual state with the desired state.

Node (worker) node

Worker nodes run the containerized applications and services, executing the tasks assigned by the master.

Key components:

kubelet – an agent on each node that receives pod specifications from the master, manages container lifecycles, and reports status.

kube-proxy – maintains network rules to enable communication between pods and provides service load balancing.

Container runtime (e.g., Docker, containerd) – creates and manages containers.

Kubernetes workflow

Users submit resource creation requests via kubectl or the API, which first reach the kube-apiserver. The API server validates the request and updates etcd. The kube-scheduler watches for unscheduled Pods, selects a suitable node, and binds it. The kube-controller-manager monitors the system, ensuring replicas and desired states. The target node’s kubelet launches the containers and continuously reports status to the master, while kube-proxy ensures network access and load balancing.