Kubernetes Complete Reference

A 10-page reference for Kubernetes operators, developers, and SREs.

Chapter 1: Core kubectl Commands

Command Description

kubectl get pods -A List all pods in all namespaces

kubectl get nodes List cluster nodes

kubectl describe pod <name> Detailed pod info

kubectl logs <pod> -f Stream pod logs

kubectl logs <pod> -c <container> Logs from specific container

kubectl exec -it <pod> -- bash Shell into a pod

kubectl apply -f [Link] Apply a manifest

kubectl delete -f [Link] Delete resources from manifest

kubectl diff -f [Link] Preview changes before apply

kubectl rollout status deploy/x Check rollout status

kubectl rollout undo deploy/x Rollback a deployment

kubectl scale deploy/x --replicas=3 Scale a deployment

kubectl port-forward svc/x 8080:80 Port-forward a service

Chapter 2: Namespaces & Contexts
Command Description

kubectl get ns List namespaces

kubectl create ns dev Create namespace

kubectl config get-contexts List kubeconfig contexts

kubectl config use-context prod Switch context

kubectl config current-context Show current context

kubectl config set-context --current --namespace=dev Set default namespace

kubectx <context> Switch context (plugin)

kubens <namespace> Switch namespace (plugin)

Chapter 2b: Resource Types Overview

Resource Abbr. Purpose

Pod po Smallest deployable unit

Deployment deploy Manages ReplicaSets & rolling updates

StatefulSet sts Stateful apps with stable IDs

DaemonSet ds One pod per node

Job job Run-to-completion tasks

CronJob cj Scheduled jobs

Service svc Stable network endpoint

Ingress ing HTTP/S routing

ConfigMap cm Non-secret config

Secret - Sensitive config (base64)

PVC pvc Storage request

HPA hpa Auto-scale on CPU/memory

Chapter 3: Pod Design & Lifecycle
Pod Phases
Phase Description

Pending Pod accepted but containers not yet created. Waiting for scheduler or image pull.

Running At least one container is running or starting/restarting.

Succeeded All containers terminated with exit code 0.

Failed At least one container terminated with non-zero exit code.

Unknown Pod state cannot be determined (usually node communication issue).

Container States
State Description

Waiting Not yet running. Reason: ContainerCreating, ImagePullBackOff, etc.

Running Container executing normally.

Terminated Container finished or crashed. Check exit code.

Common Crash Reasons

Reason Cause & Fix

CrashLoopBackOff App exits immediately. Check logs with kubectl logs.

OOMKilled Exceeded memory limit. Increase [Link].

ImagePullBackOff Image not found or no pull secret. Check image name/tag.

Pending (no nodes) Insufficient resources or taint/toleration mismatch.

Chapter 4: Services & Networking
Service Type Description Use Case

ClusterIP Internal cluster IP only Inter-service communication

NodePort Exposes on each node's IP:port Dev/testing access

LoadBalancer Cloud LB provisioned Production external access

ExternalName CNAME alias to external DNS Bridging external services

Headless No cluster IP, DNS per pod StatefulSet, direct pod DNS

Ingress Annotations (NGINX)

Annotation Purpose

[Link]/rewrite-target URL rewrite rule

[Link]/ssl-redirect Force HTTPS redirect

[Link]/proxy-body-size Max upload size

[Link]/rate-limit Rate limiting

[Link]/whitelist-source-range IP allowlist

Network Policies
Network Policies restrict pod-to-pod communication. Key concepts: podSelector selects target pods;
ingress/egress rules define allowed traffic; an empty podSelector matches all pods in the namespace.
Chapter 5: Storage
Concept Description

PersistentVolume (PV) Cluster-level storage resource provisioned by admin

PersistentVolumeClaim (PVC) User request for storage (size, access mode)

StorageClass Dynamic provisioner config (AWS EBS, GCE PD, NFS)

VolumeMount Mount a volume into a container path

emptyDir Ephemeral volume, lives with the pod

hostPath Mount node filesystem path into pod

ConfigMap volume Mount ConfigMap keys as files

Secret volume Mount Secret keys as files (in-memory tmpfs)

Access Modes
Mode Short Description

ReadWriteOnce RWO Mounted read-write by a single node

ReadOnlyMany ROX Mounted read-only by many nodes

ReadWriteMany RWX Mounted read-write by many nodes

ReadWriteOncePod RWOP Mounted by a single pod only (K8s 1.22+)

Chapter 6: Resource Management & HPA
Every container should define resource requests and limits to ensure stable scheduling and prevent
noisy-neighbor issues.

Field Description Example

[Link] Guaranteed CPU for scheduling 250m (0.25 cores)

[Link] Guaranteed memory 256Mi

[Link] CPU cap (throttled if exceeded) 500m

[Link] Memory cap (OOMKilled if exceeded) 512Mi

QoS Classes
Class Condition Eviction Priority

Guaranteed requests == limits for all containers Last to evict

Burstable requests < limits for at least one Middle priority

BestEffort No requests or limits set First to evict

HPA Configuration Reference

Field Description

minReplicas Minimum number of replicas

maxReplicas Maximum number of replicas

targetCPUUtilizationPercentage Scale out when CPU exceeds this %

[Link] Cooldown before scaling up

[Link] Cooldown before scaling down (default 300s)

Chapter 7: Probes & Health Checks
Probe Purpose Failure Action

livenessProbe Is the container alive? Restart container

readinessProbe Is the container ready for traffic? Remove from Service endpoints

startupProbe Has the container started? (slow apps) Restart if startup threshold hit

Probe Types
Type How it works

httpGet HTTP GET request; success if 200-399 returned

tcpSocket TCP connection attempt; success if port open

exec Runs command inside container; success if exit code 0

grpc gRPC Health Check Protocol (K8s 1.24+)

Key Probe Parameters

Parameter Description Default

initialDelaySeconds Wait before first probe 0

periodSeconds How often to probe 10

timeoutSeconds Probe timeout 1

failureThreshold Failures before action 3

successThreshold Successes to consider healthy 1

Chapter 8: RBAC
Role-Based Access Control (RBAC) governs what users and service accounts can do in a cluster.

Object Scope Description

Role Namespace Grants permissions within a namespace

ClusterRole Cluster-wide Grants permissions across all namespaces

RoleBinding Namespace Binds a Role to a user/group/SA

ClusterRoleBinding Cluster-wide Binds a ClusterRole cluster-wide

ServiceAccount Namespace Identity for pods to call the API

Common RBAC Verbs

Verb Description

get Read a single resource

list Read a collection

watch Stream changes

create Create a resource

update Modify an existing resource (full replace)

patch Partially modify a resource

delete Delete a resource

deletecollection Delete multiple resources

Chapter 9: Helm Package Manager
Command Description

helm repo add stable [Link] Add a chart repository

helm repo update Fetch latest chart index

helm search repo nginx Search for charts

helm install myapp ./chart Install a release

helm install myapp ./chart -f [Link] Install with custom values

helm upgrade myapp ./chart Upgrade a release

helm rollback myapp 1 Rollback to revision 1

helm list List releases

helm status myapp Release status

helm uninstall myapp Delete a release

helm template ./chart Render templates locally

helm lint ./chart Validate chart structure

helm package ./chart Package chart as .tgz

Chapter 10: Troubleshooting & Debugging
Scenario Commands to Run

Pod stuck in Pending kubectl describe pod — check Events for scheduler errors, resource limits, taint issues

Pod in CrashLoopBackOff kubectl logs <pod> --previous; check exit code in describe

Service not reachable kubectl get endpoints; verify selector matches pod labels

Image pull failure kubectl describe pod; check imagePullSecrets and registry access

Node NotReady kubectl describe node; check kubelet and container runtime status

PVC stuck in Pending kubectl describe pvc; check StorageClass and provisioner logs

High CPU/Memory kubectl top pods -A; check HPA and resource limits

Network policy blocking traffic kubectl describe networkpolicy; test with temporary allow-all policy

Debugging Toolkit
Tool Purpose

kubectl debug Ephemeral debug container in pod

kubectl run tmp --image=busybox -it Temporary

--rm debug pod

stern <pod-name> Multi-pod log tailing (plugin)

k9s Terminal-based cluster UI

kube-capacity Node/pod resource summary