Kubernetes Vocabulary: English Terms for Cloud-Native Developers

A set of machines (nodes) that run containerised workloads managed by Kubernetes. A cluster has a control plane and one or more worker nodes.

“We moved our staging environment to a separate cluster to isolate it from production.”

A physical or virtual machine in a cluster. Worker nodes run pods. The control plane manages the cluster state.

“One node is under memory pressure. We need to scale up or evict some pods.”

The set of components that manage the cluster: the API server, scheduler, controller manager, and etcd. The control plane makes global decisions about the cluster.

“The API server in the control plane is the entry point for all kubectl commands.”

A node that runs application pods. Each worker node runs the kubelet, a container runtime, and kube-proxy.

“Add a worker node to the cluster to handle the increased pod count.”

The distributed key-value store that Kubernetes uses to persist all cluster state. If etcd is lost, the cluster state is lost.

“Back up etcd daily. It is the source of truth for everything in the cluster.”

The agent that runs on each worker node. It receives pod specs from the API server and ensures the containers described in those specs are running and healthy.

“The kubelet reported the pod as CrashLoopBackOff. Check the container logs.”

The control plane component that assigns pods to nodes based on resource requirements, affinity rules, and available capacity.

“The scheduler could not place the pod because no node had enough memory.”

A virtual partition inside a cluster that isolates resources. Teams use namespaces to separate environments (dev, staging, prod) or teams.

“Deploy the app to the staging namespace to test before promoting to production.”

The smallest deployable unit in Kubernetes. A pod contains one or more containers that share the same network and storage. Most pods run a single container.

“The pod is in Pending state because there is no node with enough CPU.”

A Kubernetes resource that manages a set of identical pods. You define the desired replica count and image; Kubernetes keeps that state running and handles rolling updates.

“Update the image tag in the deployment and Kubernetes will roll it out without downtime.”

A resource that ensures a specified number of pod replicas are running at all times. Deployments manage replica sets automatically.

“The replica set scaled down to 0 pods during the maintenance window.”

Like a deployment but for stateful applications. Each pod gets a stable hostname and persistent storage. Used for databases and message queues.

“We run PostgreSQL as a StatefulSet so each pod keeps its own persistent volume.”

Ensures that one copy of a pod runs on every node (or a subset of nodes). Used for infrastructure tools like log collectors and monitoring agents.

“The log shipper runs as a DaemonSet so every node forwards its logs.”

A strategy for updating pods where new pods are started before old ones are stopped, maintaining availability throughout the update.

“The rolling update replaced pods one by one with zero downtime.”

A release strategy that sends a small percentage of traffic to the new version before rolling it out to all users.

“We ran a canary deployment with 5% of traffic to catch regressions before the full rollout.”

A release strategy that maintains two identical environments (blue and green). Traffic is switched from one to the other in a single step, enabling instant rollback.

“Blue-green gave us a one-second cutover and instant rollback when we found a bug.”

A Kubernetes resource that exposes a set of pods as a stable network endpoint with a fixed IP and DNS name. Traffic is load-balanced across healthy pods.

“Create a ClusterIP service so other pods can reach the API by name.”

A resource that routes external HTTP and HTTPS traffic into the cluster, forwarding requests to services based on hostname or URL path.

“Add an ingress rule to route /api to the backend service and / to the frontend.”

A key-value pair attached to any Kubernetes resource for identification and grouping. Selectors use labels to find matching resources.

“Add the label environment: production to all resources in the prod namespace.”

A query that matches resources by their labels. Services and deployments use selectors to target the correct pods.

“The service selector app: frontend targets all pods with that label.”

Arbitrary metadata attached to a Kubernetes resource. Unlike labels, annotations are not used for selection but for tooling, auditing, or documentation.

“We use annotations to store the deployment timestamp and the commit SHA.”

A Kubernetes resource for storing non-sensitive configuration data as key-value pairs. ConfigMaps can be mounted as files or injected as environment variables.

“Store the feature flags in a ConfigMap so we can change them without rebuilding the image.”

Like a ConfigMap but for sensitive data such as passwords, tokens, and certificates. Secrets are base64-encoded and can be encrypted at rest.

“Mount the database credentials as a secret, not as plain environment variables.”

A directory available to containers in a pod. Kubernetes volumes can be backed by many types of storage: local disk, NFS, cloud block storage, and more.

“Mount an emptyDir volume for shared temporary storage between the two containers in the pod.”

A request for storage. The cluster provisions a persistent volume to fulfil the claim. The application mounts the PVC and the data survives pod restarts.

“The database pod mounts a 50 GB PVC backed by an SSD storage class.”

The maximum CPU and memory a container is allowed to use. If it exceeds the memory limit, it is killed (OOMKilled). CPU is throttled, not killed.

“The pod was OOMKilled because the memory limit was set too low for the dataset size.”

A health check that tells Kubernetes whether the container is running correctly. If it fails, Kubernetes restarts the container.

“The liveness probe pings /healthz every 10 seconds. Three failures trigger a restart.”

A health check that tells Kubernetes whether the container is ready to receive traffic. A pod that fails the readiness probe is removed from the service's endpoints.

“Set a readiness probe so the pod does not receive traffic until the database connection is established.”

A package of pre-configured Kubernetes resources managed by Helm, the Kubernetes package manager. Charts make it easy to install, upgrade, and share applications.

“We use a Helm chart to deploy Prometheus so we can override values per environment.”

A Kubernetes extension that automates the management of complex stateful applications by encoding operational knowledge into custom controllers.

“The Postgres operator handles backups, failover, and schema migrations automatically.”