Installation & Setup

Everything runs on minikube. You do not need a cloud account or a production cluster.

Prerequisites

Tool	Minimum Version	Install
Minikube	v1.32+	brew install minikube or docs
kubectl	v1.28+	brew install kubectl
Helm	v3.13+	brew install helm
Docker or Podman	Latest	brew install podman

Some labs have additional requirements noted in their own page.

Optional but recommended

Tool	Purpose	Install
Task	Run labs with task demo -- simple-app	brew install go-task
jq	Parse JSON output from kubectl	brew install jq
k9s	Terminal-based cluster UI	brew install k9s

Minimum Resources

Resource	Minimum	Recommended
CPU cores	4	6+
RAM	8 GB	12+ GB
Disk	20 GB	40 GB

You do not need all resources at once. Each lab is self-contained. Clean up one lab before starting another if your machine is constrained.

Start Minikube

minikube start \
  --cpus=4 \
  --memory=8192 \
  --disk-size=40g \
  --driver=docker

If you use Podman, pass --driver=podman instead.

For heavier labs (Prometheus, CloudNativePG, Redis, real-world demos), consider allocating more memory:

minikube start --cpus=6 --memory=12288 --disk-size=40g

Verify the cluster

kubectl cluster-info
kubectl get nodes

You should see a single node in Ready state.

Enable Required Addons

Most labs benefit from these addons. Enable them once after starting minikube:

minikube addons enable ingress
minikube addons enable metrics-server

The ingress addon provides an nginx ingress controller. The metrics-server addon is required for the HPA lab and any lab that uses kubectl top.

Accessing Services

Minikube provides several ways to reach services running inside the cluster:

# Option 1: minikube service (opens browser)
minikube service <service-name> -n <namespace>

# Option 2: port-forward
kubectl port-forward svc/<service-name> <local-port>:<service-port> -n <namespace>

# Option 3: minikube tunnel (for LoadBalancer services)
minikube tunnel

Clone the Repository

git clone https://github.com/savitojs/k8s-learn-by-doing.git
cd k8s-learn-by-doing

All lab manifests live in the demos/ directory. Each lab is self-contained.

Using Taskfile (Optional)

If you installed Task, you can run labs with a single command:

# Deploy a lab
task demo -- simple-app

# Clean up a lab
task clean -- simple-app

# List all labs
task list

# Check prerequisites
task prereqs

The -- separator is required because Task passes everything after it as CLI_ARGS.

You can also run labs manually with kubectl. Every lab README shows the raw kubectl commands.

Real-World Labs

The real-world labs run multi-service architectures (Kafka, EFK, microservices, service mesh). They consume significantly more resources than the concept labs.

Before running a real-world lab:

1. Clean up any running labs first: task clean:all
2. Allocate at least 6 CPUs and 12GB RAM to minikube
3. Check each lab’s page for its specific resource requirements
4. Only run one real-world lab at a time unless you have 16GB+ RAM available

Cleanup

When you are done with all labs:

minikube delete

To clean up a single lab, follow the cleanup section on that lab’s page.

Next Step

Ready to go? Head to Your First Lab and deploy your first application.