README-amazon-eks.md

Deployment to Amazon EKS

Both the simulator and visualizer projects share the same Kubernetes cluster so the following step is done once for both projects.

Go to Getting started with eksctl for directions on setting up EKS and Kubernetes CLI tools.

Recommend that you create an EKS cluster with two or more Kubernetes nodes.

Once the Kubernetes cluster has been created you can use the following command to add it to your local Kubernetes configuration settings.

aws eks update-kubeconfig --name <your-cluster-name> --region <your-aws-region>

Some Additional CLI Setup

The kubectl CLI provides a nice Kubectl Autocomplete feature for bash and zsh. See the kubectl Cheat Sheet for instructions.

Also, consider installing kubectx, which also includes kubens. Mac:

brew install kubectx

Arch Linux:

yay kubectx

You may also want to create an alias for the kubectl command, such as kc.

alias kc=kubectl

Deploy the Kubernetes Dashboard (Optional)

You may want to deploy the Kubernetes dashboard. This is an optional step. To deploy the dashboard follow the Tutorial: Deploy the Kubernetes Dashboard (web UI).

Clone the GitHub Repo

Git clone the project repoisitory.

git clone https://github.com/mckeeh3/akka-cluster-kubernetes-visualizer.git

Build and Upload the Docker Image

After cloning the project, cd into the project directory then build the Docker image.

From the akka-cluster-kubernetes-visualizer project directory.

mvn clean package

Next, tag and push the image to the public Docker repo. This requires that you have a Docker account.

$ docker tag akka-k8-visualizer <your-docker-username>/akka-k8-visualizer:latest

...

$ docker push <your-docker-username>/akka-k8-visualizer

Create the Kubernetes namespace

Create the Kubernetes namespace. The namespace only needs to be created once.

kubectl create namespace visualizer

Set this namespace as the default for subsequent kubectl commands.

kubectl config set-context --current --namespace=visualizer

Deploy to Kubernetes

First, edit the kubernetes/akka-cluster-amazon-eks.yml file changing the Docker username.

image: <your-docker-username>/akka-k8-visualizer:latest

Then deploy the microservice.

kubectl apply -f kubernetes/akka-cluster-amazon-eks.yml
namespace/visualizer created
deployment.apps/visualizer created
role.rbac.authorization.k8s.io/pod-reader created
rolebinding.rbac.authorization.k8s.io/read-pods created

Check if the pods are running. This may take a few moments.

$ kubectl get pods -n visualizer
NAME                         READY   STATUS    RESTARTS   AGE
visualizer-5b4cf87d4d-5b7w6   1/1     Running   0          3m31s
visualizer-5b4cf87d4d-hzp7n   1/1     Running   0          3m31s
visualizer-5b4cf87d4d-pbst2   1/1     Running   0          3m31s

Scale the Akka cluster

Use the following commands to scale the Akka cluster nodes by adjusting the number of Kubernetes pods.

kubectl scale --replicas 5 deployment/visualizer -n visualizer

Enable External Access

Create a load balancer to enable access to the visualizer microservice HTTP endpoint.

$ kubectl expose deployment visualizer --type=LoadBalancer --name=visualizer-service -n visualizer

service/visualizer-service exposed

$ kubectl get services visualizer-service -n visualizer
NAME                 TYPE           CLUSTER-IP     EXTERNAL-IP                                                               PORT(S)                                        AGE
visualizer-service   LoadBalancer   10.100.91.15   ad110f7c353d94e9c8642b36206a74a8-1760520412.us-east-1.elb.amazonaws.com   8080:31961/TCP,8558:32054/TCP,2552:31715/TCP   62s

Access the visualizer web page

Use the hostname provided EXTERNAL-IP column in the above command. In this example, the host name is ad110f7c353d94e9c8642b36206a74a8-1760520412.us-east-1.elb.amazonaws.com. The URL to the visualizer as follows:

http://<EXTERNAL-IP>:8080

e.g.
http://ad110f7c353d94e9c8642b36206a74a8-1760520412.us-east-1.elb.amazonaws.com:8080