Kubernetes

Kubernetes vocabulary.

A practical reference for the cluster components, commands, and resources used to run workloads.

Architecture

Control plane to node.

The control plane makes global decisions and reconciles the cluster; worker nodes run Pods through the kubelet, network proxy, and container runtime.

Diagram: Kubernetes cluster architecture documentation ↗

Components

Control plane

The Kubernetes management layer that exposes the API, makes scheduling decisions, and continuously reconciles cluster state.

API server

The authenticated HTTP API entry point for Kubernetes objects and operations.

etcd

The strongly consistent key-value store that holds Kubernetes configuration and current cluster state.

Scheduler

The control-plane component that selects a suitable node for each unscheduled Pod.

Controller manager

A collection of controllers that observe objects and move the cluster toward the declared desired state.

Kubelet

The node agent that ensures the containers described by assigned Pods are running and healthy.

Container runtime

The software on a node that pulls images and runs containers through the Kubernetes runtime interface.

Commands

kubectl get

Lists Kubernetes resources and their current high-level status.

kubectl get pods -n production

kubectl describe

Shows detailed fields and recent events for a resource, useful for diagnosing scheduling and startup issues.

kubectl describe pod api-7d9f6d8b7f-x2k4m

kubectl apply

Creates or updates resources from declarative YAML or JSON manifests.

kubectl apply -f deployment.yaml

kubectl delete

Deletes named resources or resources selected by a manifest, label, or namespace.

kubectl delete -f deployment.yaml

kubectl logs

Reads the stdout and stderr emitted by a container in a Pod.

kubectl logs deploy/api -f

kubectl exec

Runs a command inside a running container for inspection or controlled troubleshooting.

kubectl exec -it deploy/api -- sh

kubectl rollout

Inspects, pauses, resumes, or rolls back a workload rollout.

kubectl rollout status deploy/api

Resources

Pod

The smallest deployable Kubernetes unit: one or more containers sharing networking and storage context.

Use case

Run one application instance or a tightly coupled group of containers.

Deployment

Manages a replicated, replaceable set of Pods and coordinates rolling updates and rollbacks.

Use case

Run stateless web servers, APIs, and workers with controlled releases.

DaemonSet

Ensures a copy of a Pod runs on every eligible node, or on a selected set of nodes.

Use case

Run node-level agents such as log collectors, monitoring agents, or storage plugins.

StatefulSet

Manages stateful Pods with stable identities, ordered deployment, and persistent storage association.

Use case

Run databases, queues, and clustered systems that need stable names or volumes.

Service

Provides a stable virtual endpoint and load balancing for a changing set of Pods.

Use case

Let clients reach a workload without knowing which Pods are currently running.

Ingress

Declares HTTP or HTTPS routing from outside the cluster to Services through an ingress controller.

Use case

Expose multiple web applications behind shared host, path, and TLS rules.

ConfigMap

Stores non-sensitive configuration that can be injected into Pods as environment variables or files.

Use case

Separate environment-specific settings from an application image.

Secret

Stores sensitive values for workloads, though encryption at rest, access control, and rotation still require deliberate configuration.

Use case

Provide credentials, tokens, certificates, or connection strings to workloads.

Namespace

Provides a logical scope for names, access policies, quotas, and environment or team boundaries.

Use case

Separate teams or environments and apply distinct access and resource policies.

PersistentVolumeClaim

A workload’s request for durable storage that Kubernetes binds to an available PersistentVolume.

Use case

Give stateful workloads storage that survives Pod replacement.

CustomResourceDefinition

Extends the Kubernetes API with a new resource kind that operators or controllers can reconcile.

Use case

Model platform-specific APIs and let an operator manage their lifecycle.

Service types

ClusterIP

The default Service type with an internal virtual IP reachable from within the cluster.

Use case

Expose an internal API or database to other workloads.

NodePort

Opens a port on each node and forwards traffic to the Service’s selected Pods.

Use case

Provide simple external access when a cloud load balancer is unavailable or unnecessary.

LoadBalancer

Requests an external load balancer from the infrastructure or cloud provider and routes traffic to the Service.

Use case

Publish a production application through a provider-managed external endpoint.

ExternalName

Maps a Service name to an external DNS name instead of selecting Pods inside the cluster.

Use case

Give applications a stable Kubernetes name for an external database or managed service.

Headless Service

A Service with `clusterIP: None` that returns Pod addresses directly through DNS.

Use case

Enable client-side discovery for StatefulSets and systems that manage their own membership.