Spring Cloud: Deployment in GKE

Posted on April 10, 2024 • 4 minutes • 655 words • Other languages: Español

This is a continuation of Spring Cloud: api-gateway and naming-server concepts .

Check the code!

You can check the code in the following repos (in all of them, stick to the branch feature/gke. You may find other branches, that’s me experimenting other solutions).

Always remember to delete your cluster when going away

DELETE THE CLUSTER! DELETE THE CLUSTER! Don’t forget of deleting your cluster when going away. Clusters burn money for just existing. Is not that expensive, but, is money going to waste.

When writing this blog, I decided to go to sleep and continue the following day, leaving a cluster running for almost 10 hours.

spending

Lucky me, cluster was empty, so the loss is negible (only £0.016). Could’ve been worse.
You can see on April 4th I spent almost a whole £1. On that day, I left a cluster with 4 microservices running for about 5 hours.
On April 8th I did the same but with Cloud Run, another Google Cloud computing solution, that deals with docker images but is less ops oriented. Because the spendings there are pay-per-use, you can see is much more afforable. Will probably do a Cloud Run blog next.
My Cloud Console is in Spanish, and my spendings are in British Pound (£) even though I’m living in Portugal. What a mess is everything lol.

Create a GKE Cluster

I asume you have everything in your Google Console ready to go. If not, Google is your best friend here.

A cluster can be created in various ways. I decided to go to ChatGPT, explained what I wanted to deploy, to keep it cheap, and came out with this glocud cli command:

gcloud beta container --project "fujiwara-383901" clusters create "pollito-demo-cluster" --no-enable-basic-auth --cluster-version "1.27.8-gke.1067004" --release-channel "regular" --machine-type "e2-medium" --image-type "COS_CONTAINERD" --disk-type "pd-ssd" --disk-size "50" --metadata disable-legacy-endpoints=true --scopes "https://www.googleapis.com/auth/devstorage.read_only","https://www.googleapis.com/auth/logging.write","https://www.googleapis.com/auth/monitoring","https://www.googleapis.com/auth/servicecontrol","https://www.googleapis.com/auth/service.management.readonly","https://www.googleapis.com/auth/trace.append" --num-nodes "3" --logging=SYSTEM,WORKLOAD --monitoring=SYSTEM --enable-ip-alias --network "projects/fujiwara-383901/global/networks/default" --subnetwork "projects/fujiwara-383901/regions/europe-southwest1/subnetworks/default" --no-enable-intra-node-visibility --default-max-pods-per-node "110" --security-posture=standard --workload-vulnerability-scanning=disabled --no-enable-master-authorized-networks --addons HorizontalPodAutoscaling,HttpLoadBalancing,GcePersistentDiskCsiDriver --enable-autoupgrade --enable-autorepair --max-surge-upgrade 1 --max-unavailable-upgrade 0 --enable-autoprovisioning --min-cpu 1 --max-cpu 2 --min-memory 2 --max-memory 6 --enable-autoprovisioning-autorepair --enable-autoprovisioning-autoupgrade --autoprovisioning-max-surge-upgrade 1 --autoprovisioning-max-unavailable-upgrade 0 --autoscaling-profile optimize-utilization --enable-vertical-pod-autoscaling --enable-shielded-nodes --zone "europe-southwest1-a"

Adapt to your project and preferred region.

Deploy stuff in the cluster

The order in which the microservices are deployed doesn’t really matter, but to prevent unnecessary errors in the logs, I follow this order:

naming-server
api-gateway
microservice-b
microservice-a

Get gcloud CLI and kubectl in your machine.
Click on “Establish Connection”. That will give us a gcloud command.

establish connection

Paste the command in your fav cmd tool.

paste

Navigate to where deployment.yaml is. I decided to put the file in deployment/kubernetes/prod
Run

kubectl apply -f deployment.yaml

You should get something like this:

E0411 12:36:32.181316   19120 memcache.go:287] couldn't get resource list for metrics.k8s.io/v1beta1: the server is currently unable to handle the request
E0411 12:36:32.280107   19120 memcache.go:121] couldn't get resource list for metrics.k8s.io/v1beta1: the server is currently unable to handle the request
Warning: autopilot-default-resources-mutator:Autopilot updated Deployment default/naming-server: adjusted resources to meet requirements for containers [spring-cloud-v2-naming-server] (see http://g.co/gke/autopilot-resources)
deployment.apps/naming-server created
service/naming-server created

Now, let’s repeat the deployment process for the rest of the microservices.

After some minutes, run:

kubectl get pods

to check that everything is running smoothly.

NAME                              READY   STATUS    RESTARTS   AGE
api-gateway-85cd954d8-57kp2       1/1     Running   0          10m
microservice-a-6978854d7b-24j45   1/1     Running   0          58m
microservice-b-6f6c5944b4-tgcvb   1/1     Running   0          17m
naming-server-f4795fd84-q9l8x     1/1     Running   0          14m

Let’s test it

To get the url, run:

kubectl get svc

NAME             TYPE           CLUSTER-IP       EXTERNAL-IP    PORT(S)          AGE
api-gateway      LoadBalancer   10.111.196.131   34.175.79.29   8765:30836/TCP   114m
kubernetes       ClusterIP      10.111.192.1     <none>         443/TCP          126m
microservice-a   ClusterIP      10.111.196.50    <none>         8081/TCP         63m
microservice-b   ClusterIP      10.111.194.149   <none>         8080/TCP         110m
naming-server    ClusterIP      10.111.193.218   <none>         8761/TCP         116m

You’ll see only one has an EXTERNAL-IP, the api-gateway. Grab that url, add the port at the end, add /microservice-a, and we should get a result back.

curl --location '34.175.79.29:8765/microservice-a'

postman

Congrats, you now deployed your stuff in GKE.

Don’t forget to delete the cluster when you are done

Reminder :D

Next step

Deploy it in Cloud Run, cause why not.