This demo uses a slightly modified version of the HorizontalPodAutoscaler Walkthrough example from the Kuberentes documentation. It creates the following resources:

• hpa-demo-app Deployment
• hpa-demo-service Service
• hpa-demo HorizontalPodAutoscaler with minReplicas=5, maxReplicas=40 and targetCPUUtilizationPercentage=70

To create a cluster, run

gcloud beta container clusters \
create-auto hpa-demo --region us-central1 \
--cluster-version "1.32.1-gke.1729000" \
--release-channel "rapid" \
--hpa-profile performance

To deploy, run

kubectl apply -f manifests

For generating load, the load.sh script is provided. It uses the simple load test application hey to generate load for 2 minutes with 10 concurrent clients. You can adjust the parameters in the script as you see fit.