15_HPA_Lab

K8s | HPA Lab

In this lab we will use HPA to ensure that k8s will take care of scaling and de-scaling of pods of your MS on the basis of CPU & Memory usage :

• Enable autoscaling for Attendance MS with a threshold of 50% avg CPU usage of all instances.

By the end of this lab, you will have to be able to deploy your microservices in such a fashion that they can scale up & down on the basis of CPU/memory usages.

Pre-requisities

We have to restart minikube to enable some config for HPA

minikube start --extra-config=controller-manager.horizontal-pod-autoscaler-upscale-delay=1m --extra-config=controller-manager.horizontal-pod-autoscaler-downscale-delay=1m --extra-config=controller-manager.horizontal-pod-autoscaler-sync-period=10s --extra-config=controller-manager.horizontal-pod-autoscaler-downscale-stabilization=1m

minikube addons enable metrics-server

Application

So firstly, we have to give resources and limits because HPA uses them to calculate usage and then scale it.

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: php-apache-<name>
spec:
  selector:
    matchLabels:
      run: php-apache-<name>
  replicas: 1
  template:
    metadata:
      labels:
        run: php-apache-<name>
    spec:
      containers:
      - name: php-apache
        image: k8s.gcr.io/hpa-example
        ports:
        - containerPort: 80
        resources:
          limits:
            cpu: 500m
          requests:
            cpu: 200m

kubectl apply -f hpa.yaml

---
apiVersion: v1
kind: Service
metadata:
  name: php-apache-<name>
  labels:
    run: php-apache-<name>
spec:
  ports:
  - port: 80
  selector:
    run: php-apache

kubectl apply -f hpa-service.yaml

Create an HPA manifest for attendance application

apiVersion: autoscaling/v1
kind: HorizontalPodAutoscaler
metadata:
  name: php-apache-<name>
spec:
  maxReplicas: 10
  minReplicas: 1
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: php-apache-<name>
  targetCPUUtilizationPercentage: 50

kubectl apply -f php-hpa.yaml

Validate it by

kubectl get hpa -w

Generate Load and Testing

Open a new tab and run a pod with the name load-generator

kubectl run -it --rm load-generator --image=busybox /bin/sh

Run this command:-

while true; do wget -q -O- http://php-apache-<name>; done

And wait for 2 to 3 minutes and you will have something like this:-