Kubernetes Reference Guide

Kubernetes Glossary

A plain-language glossary of the Kubernetes terms used throughout the training path.

How to use this page

Kubernetes has a lot of terms. This glossary gives short practical definitions for the objects and concepts used throughout the training path.

Cluster and node terms

Cluster

A Kubernetes environment made up of a control plane and worker nodes.

Node

A machine that runs Pods. In kind, nodes run as Docker containers. In production, nodes are usually VMs or cloud instances.

Control plane

The Kubernetes components that manage cluster state, scheduling, API requests, controllers, and coordination.

API server

The front door to Kubernetes. kubectl and controllers communicate with the Kubernetes API server.

kubelet

The node agent that makes sure containers are running as expected on a node.

Workload terms

Pod

The smallest deployable unit Kubernetes manages. A Pod wraps one or more containers.

Container

A running process created from a container image.

Image

A packaged application and runtime filesystem used to create containers.

Deployment

A controller that manages stateless application Pods, rollouts, scaling, and rollback behavior.

ReplicaSet

A controller that maintains a desired number of matching Pods. Deployments create and manage ReplicaSets.

StatefulSet

A controller for stateful workloads that need stable identity and storage patterns.

DaemonSet

A controller that runs a copy of a Pod on selected nodes.

Job

A workload that runs to completion.

CronJob

A workload that creates Jobs on a schedule.

Networking terms

Service

A stable network abstraction that routes traffic to matching Pods.

ClusterIP

The default Service type. It exposes a stable IP inside the cluster.

NodePort

A Service type that exposes a port on cluster nodes.

LoadBalancer

A Service type that requests an external load balancer from the infrastructure provider.

Endpoint

A backend network destination behind a Service, usually a Pod IP and port.

Ingress

A Kubernetes object that defines HTTP and HTTPS routing rules into Services.

Ingress controller

The software that watches Ingress resources and actually routes incoming traffic.

Configuration and storage terms

ConfigMap

A Kubernetes object used to store non-sensitive configuration data.

Secret

A Kubernetes object used to store sensitive values such as tokens, keys, and passwords. Secret values still require careful access control.

Volume

Storage mounted into a Pod.

PersistentVolume

A storage resource available to the cluster.

PersistentVolumeClaim

A request for storage by a workload.

StorageClass

Defines how storage should be provisioned.

Security terms

ServiceAccount

An identity used by workloads running inside Kubernetes.

RBAC

Role-Based Access Control. Kubernetes authorization model for granting permissions.

Role

A set of permissions inside a Namespace.

ClusterRole

A set of permissions that can apply cluster-wide or across Namespaces.

RoleBinding

Grants a Role or ClusterRole to a subject inside a Namespace.

NetworkPolicy

A policy that controls allowed Pod network traffic when enforced by the cluster networking plugin.

Operations terms

Namespace

A logical grouping of Kubernetes resources inside a cluster.

Request

The CPU or memory amount a container asks Kubernetes to reserve for scheduling.

Limit

The maximum CPU or memory amount a container is allowed to use.

Readiness probe

A health check that controls whether a Pod should receive Service traffic.

Liveness probe

A health check that controls whether Kubernetes should restart a container.

Startup probe

A health check that gives slow-starting applications time to initialize before other probes take over.

Event

A Kubernetes record of something the platform tried to do or observed.

Rollout

The process of updating a Deployment to a new version.

Rollback

Reverting a Deployment to a previous working revision.

Scheduling and Helm terms

Scheduler

The Kubernetes component that decides which node a Pod should run on.

nodeSelector

A simple way to schedule Pods only onto nodes with matching labels.

Node affinity

A more expressive way to influence Pod placement based on node labels.

Taint

A setting on a node that repels Pods unless they tolerate it.

Toleration

A setting on a Pod that allows it to be scheduled onto a node with a matching taint.

Helm

A package manager for Kubernetes applications.

Chart

A packaged Kubernetes application.

Values

Configuration inputs used to render a Helm chart.

Release

An installed instance of a Helm chart in a cluster.

Bottom line: Kubernetes terms become easier when you connect them to the operating model: controllers manage desired state, Services provide stable access, policies define guardrails, and kubectl helps you inspect what is actually happening.
← Back to Kubernetes Field Notes Series index Final Project →