Glossary
Short, plain-language definitions of the terms used across this site. If a page uses a word you have not met, it is probably here.
Cluster basics
| Term |
Definition |
| Node |
A machine, physical or virtual, that runs your containers. |
| Cluster |
A set of nodes plus the control plane that manages them. |
| Control plane |
The components that decide what should happen: API server, scheduler, controller manager, and etcd. It does not run your application containers. |
| API server |
The single entry point for every change to the cluster. Every other component reads and writes through it. |
| etcd |
The distributed key-value store holding all cluster state. Only the API server talks to it directly. |
| kubectl |
The command-line tool you use to talk to a cluster (pronounced "kube control"). |
| kubelet |
The agent on each node that starts containers and reports their status back to the API server. |
| Manifest |
A YAML (or JSON) file describing one or more Kubernetes objects. What you feed to kubectl apply. |
Objects you run
| Term |
Definition |
| Pod |
The smallest thing Kubernetes runs: one or more containers scheduled together on one node, sharing an IP address. |
| Deployment |
The controller that keeps a set of identical Pods running and updates them safely. |
| ReplicaSet |
The object a Deployment creates to hold one specific version of a Pod template at a specific replica count. |
| StatefulSet |
Like a Deployment, but for workloads whose replicas are not interchangeable - stable names, stable storage. |
| DaemonSet |
Runs one copy of a Pod on every node (or every node matching a selector). |
| Job / CronJob |
A workload that runs to completion, once (Job) or on a schedule (CronJob). |
| Service |
A stable virtual address and DNS name in front of a changing set of Pods. |
| Ingress / Gateway API |
HTTP and TLS routing for traffic entering the cluster from outside. |
| ConfigMap / Secret |
Namespace-scoped objects holding configuration values and sensitive values respectively. |
| Namespace |
A logical partition inside a cluster. Partitions object names and policy scope - not the network. |
How it works
| Term |
Definition |
| Controller |
A loop that watches objects and takes action until the observed state matches the declared state. |
| Reconciliation |
That process of closing the gap between spec (what you want) and status (what is true). |
| spec / status |
The two halves of almost every Kubernetes object: declared intent, and observed reality. |
| Label |
An arbitrary key-value tag attached to an object, e.g. app: web. |
| Selector |
A query over labels. This is how a Service finds its Pods and a Deployment finds its ReplicaSet. |
| Admission controller |
Code that runs inside the API server after authorization but before an object is stored, able to modify it (mutating) or reject it (validating). |
| Operator / CRD |
A CustomResourceDefinition adds your own object type to the API; an operator is the controller that reconciles it. |
Security and networking acronyms
| Term |
Definition |
| RBAC |
Role-Based Access Control. The system that decides which users and workloads may read or change which objects. |
| Service account |
The identity a Pod uses when it calls the API server. |
| NetworkPolicy |
Rules restricting which Pods may talk to which. Requires a CNI that implements it. |
| PSA / Pod Security Standards |
Per-namespace limits on what a Pod is allowed to declare, such as running as root. |
| CNI |
Container Network Interface. The plugin standard Kubernetes uses to hand pod networking to an implementation such as Cilium, Calico, or Flannel. |
| CIDR |
A way of writing a range of IP addresses, like 10.244.0.0/16. The number after the slash says how many leading bits are fixed. |
| NAT |
Network Address Translation: rewriting the source or destination IP of a packet in flight. The pod network deliberately avoids it between Pods. |
| Overlay network |
Pod traffic wrapped inside another packet so it can cross a network that knows nothing about pod IPs. |
| East-west / north-south |
Traffic between workloads inside the cluster, versus traffic entering or leaving it. |
| IAM / KMS |
Your cloud provider's identity system, and its key management service (used to encrypt Secrets at rest). |
Resources and scheduling
| Term |
Definition |
| Request |
The amount of CPU or memory a container is guaranteed. The scheduler places Pods using requests, not actual usage. |
| Limit |
The ceiling a container may use. Exceeding a CPU limit throttles the process; exceeding a memory limit kills it. |
| QoS class |
Guaranteed, Burstable, or BestEffort - derived from requests and limits, and used to decide eviction order. |
| Taint / toleration |
A taint marks a node as repelling Pods; a toleration lets a specific Pod land there anyway. |
| Affinity |
Rules expressing which nodes a Pod may or prefers to run on, or which other Pods it wants to sit near or apart from. |
| PV / PVC / StorageClass |
The actual storage, the request for storage, and the template that provisions the first from the second. |