Deployment with HPA #

We will create a simple helloweb application with resource limits and hpa based on CPU usage. First create a GKE clutser and login to the cluster via kubectl.

Create the deployment with resource #

kubectl apply -f dc-myapp.yaml

Make sure you have define resources in the deployment yaml inroder to make hpa (based on cpu) to be worked.

dc-myapp.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: helloweb
  labels:
    app: hello
spec:
  selector:
    matchLabels:
      app: hello
      tier: web
  template:
    metadata:
      labels:
        app: hello
        tier: web
    spec:
      containers:
      - name: hello-app
        image: gcr.io/google-samples/hello-app:1.0
        ports:
        - containerPort: 8080
        resources:
          limits:
            cpu: 200m
            memory: 300Mi
          requests:
            cpu: 100m
            memory: 100Mi

Create LoadBalancer service for the deployment #

kubectl apply -f svc-myapp.yaml

svc-myapp.yaml

apiVersion: v1
kind: Service
metadata:
  name: helloweb
  labels:
    app: hello
spec:
  selector:
    app: hello
    tier: web
  ports:
  - port: 80
    targetPort: 8080
  type: LoadBalancer

Create HPA for the deployment #

Create autoscale when cpu hit for 50%

kubectl autoscale deployment helloweb --max 6 --min 2 --cpu-percent 50

See the hpa. This will show all the hpa objects in the namespace

kubectl get hpa

Get LoadBalancer IP #

Shows all the services in the namesapace. Identify the helloweb svc and its external IP. Using that external IP you can access the app via a web browser.

kubectl get svc

Generate load for the app #

Here we create some load to that app. This will increase the cpu and then hpa will kick in and will spin few new pods.s

ab -k -c 1000 -n 2000000 http://<IP-of-the-loadbalancer>/