Day 24: Scaling Applications with Kubernetes HPA and Cluster Autoscaler

Welcome to Day 24 of the Zero to Platform Engineer in 30 Days challenge! 🚀 Today, we’re diving into autoscaling in Kubernetes, focusing on Horizontal Pod Autoscaler (HPA) and Cluster Autoscaler to efficiently manage workloads.

Why Autoscaling Matters in Kubernetes?

Autoscaling ensures that applications:

• Adapt to demand by increasing or decreasing resources dynamically.
• Improve cost efficiency by running only what’s needed.
• Maintain performance under varying loads.

Two main types of autoscaling in Kubernetes:

• HPA (Horizontal Pod Autoscaler) → Scales pods based on CPU, memory, or custom metrics.
• Cluster Autoscaler → Adjusts the number of nodes dynamically based on pod scheduling needs.

Horizontal Pod Autoscaler (HPA)

• Scales pods automatically based on CPU, memory, or custom metrics.
• Works well for applications with fluctuating workloads.
• Requires metrics-server to be installed.

Step 1: Enable Metrics Server

kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

Step 2: Deploy a Sample Application

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx
        resources:
          requests:
            cpu: "100m`
          limits:
            cpu: "200m"

kubectl apply -f deployment.yaml

Step 3: Create an HPA for the Deployment

kubectl autoscale deployment nginx-deployment --cpu-percent=50 --min=1 --max=10

Step 4: Check HPA Status

kubectl get hpa

💡 HPA will increase/decrease pods based on CPU usage automatically.

Cluster Autoscaler

• Scales nodes up/down when pods can’t be scheduled due to resource constraints.
• Works with cloud providers like AWS, GCP, and Azure.

Step 1: Deploy Cluster Autoscaler on AWS EKS

helm repo add autoscaler https://kubernetes.github.io/autoscaler
helm repo update
helm install cluster-autoscaler autoscaler/cluster-autoscaler \
  --namespace kube-system \
  --set autoDiscovery.clusterName=my-cluster \
  --set awsRegion=us-west-2

Step 2: Verify Cluster Autoscaler

kubectl logs -f deployment/cluster-autoscaler -n kube-system

💡 Cluster Autoscaler ensures that Kubernetes adds/removes nodes as needed.

Activity for Today

1. Deploy HPA for a sample application.
2. Enable Cluster Autoscaler and scale up/down nodes.
3. Test autoscaling by generating load on your application.

What’s Next?

Tomorrow, we’ll explore safe deployments using Canary Releases and Feature Flags.

👉 Check it out here: Zero to Platform Engineer Repository

Feel free to clone the repo, experiment with the code, and even contribute if you’d like! 🚀

Follow the Series!

🎉 Don’t miss a single step in your journey to becoming a Platform Engineer! 🎉

This post is just the beginning. Here’s what we’ve covered so far and what’s coming up next:

• Day 0: Introduction - What’s Platform Engineering?

• Day 1: Introduction to the CNCF Landscape

• Day 2: Day 2: Containers and Docker - The Building Blocks of Cloud Native

• Day 3: Containers and Kubernetes - The Building Blocks of Cloud Native

• Day 4: Deployments and Scaling in Kubernetes - Let’s Get Practical

• Day 5: Kubernetes Services – Connecting Your Applications

• Day 6: ConfigMaps and Secrets – Managing Configurations in Kubernetes

• Day 7: Recap and Hands-On Challenges for Week 1

• Day 8: Introduction to Infrastructure as Code (IaC)

• Day 9: Advanced Terraform – Managing Kubernetes Resources

• Day 10: Managing Kubernetes with Helm and Terraform

• Day 11: Introduction to GitOps with ArgoCD

• Day 12: Advanced GitOps with Argo Rollouts

👉 Bookmark this blog and check back every day for new posts in the series. 📣 Share your progress on social media with the hashtag #ZeroToPlatformEngineer to connect with other readers!