Kubernetes (K8s) Cheatsheet
Kubernetes is an open-source container orchestration platform designed to automate application deployment, scaling, and management.
1. Core Kubectl Commands
# Cluster Information
kubectl cluster-info # Display cluster info
kubectl get nodes # List nodes in the cluster
kubectl describe node <node-name> # Show detailed node information
# Resource Creation & Management
kubectl apply -f manifest.yaml # Create/update resource(s) from a manifest file
kubectl delete -f manifest.yaml # Delete resource(s) defined in a manifest
kubectl delete pod <pod-name> # Delete a specific pod
2. Viewing & Inspecting Resources
# Get Commands (List resources)
kubectl get pods # List all pods in the active namespace
kubectl get pods -A # List all pods across all namespaces
kubectl get services # List services
kubectl get deployments # List deployments
kubectl get namespaces # List namespaces
# Output Formatting
kubectl get pods -o wide # List pods with more detailed columns (IP, Node)
kubectl get pod <pod-name> -o yaml # Output the pod definition in YAML format
kubectl get pods --sort-by='.metadata.creationTimestamp' # Sort list by creation date
# Describe (Deep dive)
kubectl describe pod <pod-name> # Deep inspect pod events and specifications
kubectl describe deployment <deploy-name># Deep inspect deployment
3. Debugging & Troubleshooting
# Logging
kubectl logs <pod-name> # Stream logs from a single-container pod
kubectl logs <pod-name> -c <container> # Stream logs for a specific container in a pod
kubectl logs -f <pod-name> # Follow log stream (real-time tail)
kubectl logs --tail=100 <pod-name> # Print the last 100 lines of logs
# Shell Execution
kubectl exec -it <pod-name> -- /bin/sh # Open interactive shell inside pod
kubectl exec -it <pod-name> -c <container> -- /bin/bash # Open shell inside multi-container pod
# Networking & Port Forwarding
kubectl port-forward <pod-name> 8080:80 # Forward local port 8080 to pod port 80
kubectl port-forward svc/<service> 8080:80 # Forward local port 8080 to service port 80
4. Multi-Container Pod Patterns
Kubernetes supports multiple containers sharing the same Pod network and volume namespaces.
graph TD
subgraph Pod
Main[Main Application Container] <-->|Shares localhost & volumes| Sidecar[Sidecar Container: e.g., Log Shipper]
SharedVol[(Shared EmptyDir Volume)] --> Main
SharedVol --> Sidecar
end
1. Sidecar Pattern
An auxiliary container enhances the main container (e.g., streaming logs, syncing configurations, or downloading assets).
apiVersion: v1
kind: Pod
metadata:
name: sidecar-example
spec:
containers:
- name: main-app
image: alpine
command: ["/bin/sh", "-c", "while true; do echo $(date) 'User log' >> /var/log/app.log; sleep 5; done"]
volumeMounts:
- name: log-vol
mountPath: /var/log
- name: sidecar-shipper
image: alpine
command: ["/bin/sh", "-c", "tail -f /var/log/app.log"]
volumeMounts:
- name: log-vol
mountPath: /var/log
volumes:
- name: log-vol
emptyDir: {}
5. Standard Manifest Examples
Deployment Manifest with SecurityContext
apiVersion: apps/v1
kind: Deployment
metadata:
name: frontend-deployment
labels:
app: frontend
spec:
replicas: 3
selector:
matchLabels:
app: frontend
template:
metadata:
labels:
app: frontend
spec:
securityContext:
runAsNonRoot: true
runAsUser: 1000
containers:
- name: web-app
image: nginx:alpine
securityContext:
allowPrivilegeEscalation: false
readOnlyRootFilesystem: true
ports:
- containerPort: 80
resources:
requests:
memory: "64Mi"
cpu: "250m"
limits:
memory: "128Mi"
cpu: "500m"
NetworkPolicy Specification
By default, Kubernetes pods accept traffic from any source. A NetworkPolicy isolates traffic.
apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
name: allow-frontend-only
namespace: database
spec:
podSelector:
matchLabels:
role: db
policyTypes:
- Ingress
ingress:
- from:
- podSelector:
matchLabels:
role: frontend
ports:
- protocol: TCP
port: 5432
6. Storage Provisioning (PV & PVC)
Dynamic volume provisioning lets cluster administrators define StorageClasses so that physical disks are provisioned automatically on demand.
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: pg-storage-pvc
spec:
accessModes:
- ReadWriteOnce
storageClassName: standard # Maps to cloud provider storage driver
resources:
requests:
storage: 10Gi
7. Helm Charts Usage
Helm is the package manager for Kubernetes, packaging resource manifests into standardized templates.
# Add a repository
helm repo add bitnami https://charts.bitnami.com/bitnami
helm repo update
# Install a chart
helm install my-postgres bitnami/postgresql --set auth.database=prod_db
# Upgrade an existing deployment
helm upgrade my-postgres bitnami/postgresql --values overrides.yaml
# List active releases
helm list
8. Common Gotchas & Troubleshooting Steps
CrashLoopBackOff: The container starts, crashes, restarts, and crashes again. Check logs immediately withkubectl logs <pod-name>. Often caused by incorrect commands, missing environment variables, or application errors.ImagePullBackOff: Kubernetes is unable to pull the container image. Check image name capitalization, spelling, tag/version, and ensure pull secrets are present if using a private registry.- Namespace Scope: If you cannot find your resource, you might be in the wrong namespace. Append
-n <namespace>to check, or use-Ato see everything. - Service Matching Labels: A Service forwards traffic to Pods via selector matching. If your Service is not routing traffic, verify that
spec.selectormatches the labels defined on your Pods exactly.
Best Practices & Production Standards
- Enforce Resource Limits: Always declare explicit CPU/Memory
requestsandlimitsto enable stable pod scheduling. - Configure Health Probes: Define robust
livenessProbes,readinessProbes, andstartupProbesto enable automated pod self-healing. - Enforce Namespace Isolation: Partition cluster resources into logical Namespaces, and lock down communication using NetworkPolicies.
Common Mistakes & Antipatterns
- Deploying with Hardcoded NodePorts: Exposing service endpoints with high NodePort numbers on master nodes instead of using standard Ingress resources.
- Using
:latestImage Tags: Deploying container deployments without deterministic tags, causing mismatched running image versions on pod restarts. - Mixing Config and Code: Storing secret variables in base manifests instead of deploying them as highly-typed Secrets or ConfigMaps.
Troubleshooting & Debugging Guide
- CrashLoopBackOff Diagnostics: Call
kubectl logs <pod-name>to read boot-time errors, and runkubectl describe pod <pod-name>to analyze scheduling events, volumes, and probe failures. - Inbound Routing Failures: Check services and selectors. Run
kubectl get endpointsto verify pods are mapped correctly to target services.
Core Interview Questions & Answers
- Q: Describe the role of the Kubernetes Control Plane and Worker Nodes.
- A: The Control Plane manages cluster state (API Server, etcd, Scheduler, Controller Manager). Worker Nodes run container applications and include the Kubelet agent, Kube-Proxy router, and Container Runtime (e.g., containerd).
- Q: What is the difference between a Liveness and a Readiness probe?
- A: A Liveness probe determines if a container is alive; if it fails, Kubernetes kills and restarts the pod. A Readiness probe determines if a container is ready to serve network traffic; if it fails, Kubernetes removes the pod from service endpoints.
Technical Architecture Diagram
graph TD
APIServer[Kube API Server] --> Scheduler[Kube Scheduler]
APIServer --> ControlManager[Kube Controller Manager]
APIServer --> etcd[(etcd Cluster State)]
APIServer --> Kubelet[Kubelet Node Agent]