What Kubernetes is, and the problem it solves

Kubernetes (K8s) is an open-source container orchestration platform. It automates deployment, scaling, networking, and self-healing for containerized applications — originally built at Google, now maintained by the CNCF.

Why it exists

Docker alone is great at packaging a single container — but running dozens or hundreds of them in production exposes problems it was never designed to solve.

The stack, top to bottom

Core features at a glance

Cluster architecture: control plane vs. worker nodes

Kubernetes follows a master–worker (client–server) model. Every cluster splits into two halves: the Control Plane, which makes decisions, and Worker Nodes, which do the work.

Control plane components

Worker node components

Pods: the smallest deployable unit

A Pod wraps one or more containers that share the same network and storage. Containers inside a Pod are always scheduled together, on the same node — they’re never split up.

Why a Pod instead of a bare container?

Pod lifecycle

Pod types

ReplicaSets & Deployments

You almost never create Pods directly. Instead, a Deployment manages a ReplicaSet, which manages the Pods — giving you scaling, rolling updates, and rollbacks for free.

ReplicaSet vs. Deployment

Rolling update, step by step

Rollback flow

# roll back to the previous working version
kubectl rollout undo deployment my-app
# see revision history first if you need to pick a specific one
kubectl rollout history deployment my-app

Deployment strategies

Commands you’ll actually use

Services: stable networking for ephemeral Pods

Pods are ephemeral — they get created, destroyed, and rescheduled, and their IPs change every time. A Service gives them a stable IP and DNS name, and load-balances traffic across whichever Pods currently exist.

The four Service types

Ingress: smart routing at Layer 7

A LoadBalancer or NodePort Service exposes one application. Ingress sits in front of many Services behind a single entry point, routing by hostname or path — at the HTTP layer, not just TCP/UDP.

Service vs. Ingress

# list every Ingress across all namespaces
kubectl get ingress -A
# create an Ingress from a YAML manifest
kubectl apply -f ingress.yaml
# inspect the rules on a specific Ingress
kubectl describe ingress my-ingress

ConfigMaps & Secrets

Both decouple configuration from your container image, so the same image can run in dev, test, and prod without rebuilding it. The difference is sensitivity.

Volumes, Persistent Storage & StorageClasses

Containers are ephemeral — delete one, and everything inside it is gone. Volumes give Pods storage that survives container restarts and deletions.

Volume types

How PV, PVC, and StorageClass fit together

# storage inspection
kubectl get pv
kubectl get pvc
kubectl describe pvc my-pvc
kubectl get storageclass
kubectl describe storageclass fast-ssd

Namespaces

A Namespace divides cluster resources between multiple users, teams, or environments. Resources in one namespace are invisible to another by default.

• The default namespace is literally called default.
• Namespaces help with isolation and resource management (e.g. quotas per team).
• Resources in different namespaces are isolated from each other.

# list namespaces, and everything scoped to one
kubectl get ns
kubectl get pods -n my-namespace

Resource management & health checks

Two mechanisms keep a cluster stable under real load: resource requests/limits control how much a container can use, and probes control whether it’s considered healthy.

Requests vs. Limits

The three probes

Common labels

kubectl command reference

Grouped by what you’re actually trying to do, so you can scan straight to the section you need.

Inspecting the cluster

kubectl get pods                    # list all Pods
kubectl get svc                     # list all Services
kubectl get nodes                   # list all Nodes
kubectl get ns                      # list all Namespaces
kubectl describe pod <pod-name>     # full details of a Pod
kubectl logs <pod-name>             # logs of a Pod
kubectl exec -it <pod-name> -- /bin/bash   # shell into a Pod

Deployments & ReplicaSets

kubectl create deployment <name> --image=<image>
kubectl get deployments
kubectl get rs
kubectl describe deployment <name>
kubectl scale deployment <name> --replicas=<n>
kubectl rollout status deployment <name>
kubectl rollout undo deployment <name>
kubectl rollout history deployment <name>

Services & Ingress

kubectl get svc
kubectl get ingress -A
kubectl describe ingress <name>
kubectl apply -f ingress.yaml
kubectl delete ingress <name>

Storage

kubectl get pv
kubectl get pvc
kubectl describe pvc <name>
kubectl delete pvc <name>
kubectl get storageclass
kubectl describe storageclass <name>