optimize-costs-for-google-kubernetes-engine

Optimize Costs for Google Kubernetes Engine

• Understading your GKE Costs

• Monitoring your GKE costs

• Managing a GKE Multi-tenant Cluster with Namespaces

• Virtual machines in GKE

• Exploring Cost-optimization for GKE Virtual Machines

• Autoscaling with GKE: Overview and pods

• Autoscaling with GKE: Clusters and nodes

• Understanding and Combining GKE Autoscaling Strategies

• Application optimization for GKE

• Optimize GKE to run your application

• GKE Workload Optimization

• Optimize Costs for Google Kubernetes Engine: Challenge Lab

Optimize Costs for Google Kubernetes Engine: Challenge Lab

Task 1: Create a cluster and deploy your app Bash

Set your Zone

gcloud config set compute/zone europe-west4-a

1. Create the cluster with 2 nodes and Rapid release channel

gcloud container clusters create onlineboutique-cluster-465 --project=qwiklabs-gcp-03-7c944f176028 --machine-type=e2-standard-2 --num-nodes=2 --release-channel=rapid

2. Authenticate the cluster

gcloud container clusters get-credentials onlineboutique-cluster-465 --zone europe-west4-a

3. Create the namespaces

kubectl create namespace dev kubectl create namespace prod

4. Clone and deploy the application

git clone https://github.com/GoogleCloudPlatform/microservices-demo.git cd microservices-demo kubectl apply -f ./release/kubernetes-manifests.yaml --namespace dev Task 2: Migrate to an optimized node pool Bash

1. Create the optimized node pool (Custom Machine Type)

gcloud container node-pools create optimized-pool-3828 --cluster=onlineboutique-cluster-465 --machine-type=custom-2-3584 --num-nodes=2 --zone=europe-west4-a

2. Cordon the old nodes (Mark them as unschedulable)

for node in $(kubectl get nodes -l cloud.google.com/gke-nodepool=default-pool -o=name); do kubectl cordon "$node" done

3. Drain the old nodes (Evict pods to the new pool)

for node in $(kubectl get nodes -l cloud.google.com/gke-nodepool=default-pool -o=name); do kubectl drain "$node" --force --ignore-daemonsets --delete-emptydir-data --grace-period=10 done

4. Delete the old pool

gcloud container node-pools delete default-pool --cluster=onlineboutique-cluster-465 --zone=europe-west4-a --quiet Task 3: Apply a frontend update Bash

1. Set the Pod Disruption Budget

kubectl create poddisruptionbudget onlineboutique-frontend-pdb --selector run=frontend --min-available=1 --namespace dev

2. Update the frontend image to v2.1

kubectl set image deployment/frontend server=gcr.io/qwiklabs-resources/onlineboutique-frontend:v2.1 --namespace dev

3. Set Image Pull Policy to Always

kubectl patch deployment frontend --namespace dev -p '{"spec":{"template":{"spec":{"containers":[{"name":"server","imagePullPolicy":"Always"}]}}}}' Task 4: Autoscale from estimated traffic Bash

1. Set Horizontal Pod Autoscaler for Frontend (Max 11)

kubectl autoscale deployment frontend --cpu-percent=50 --min=1 --max=11 --namespace dev

2. Enable Cluster Autoscaler (Scale nodes 1 to 6)

gcloud beta container clusters update onlineboutique-cluster-465 --enable-autoscaling --min-nodes=1 --max-nodes=6 --zone=europe-west4-a

3. Run the load test (Update with your Frontend External IP)

kubectl exec $(kubectl get pod --namespace=dev | grep 'loadgenerator' | cut -f1 -d ' ') -it --namespace=dev -- bash -c 'export USERS=8000; locust --host="http://[EXTERNAL_IP]" --headless -u "8000" 2>&1'

4. Autoscale Recommendation Service (Max 5)

kubectl autoscale deployment recommendationservice --cpu-percent=50 --min=1 --max=5 --namespace dev Verification Commands (Checkups) Bash

Check Pod status in dev

kubectl get pods -n dev

Check HPA scaling progress

kubectl get hpa -n dev

Check node scaling

kubectl get nodes

Check resource usage per pod

kubectl top pods -n dev