Provision an EKS Cluster - How to

On this page you will learn how to setup an edge system with an Amazon Kubernetes Service (EKS) cluster using the cluster installation tools in the Kelvin UI.

If you want Kelvin to be responsible to also install the Kubernetes cluster directly and handle all the management of the cluster, you can choose the K3S option.

Amazon Kubernetes Service (EKS)

Amazon Elastic Kubernetes Service (Amazon EKS) is a fully-managed service that simplifies the deployment and management of Kubernetes clusters. It provides a highly available and scalable platform to deploy containerized Kelvin SmartApps™ using Kubernetes, without the need to manage the underlying infrastructure.

Requirements & Limitations

Cloud Kubernetes implementations vary across providers, and even within the same provider, there can be different types of stacks based on the selected services.

So it is not possible to list all the requirements, limitations and fixed design options available and will heavily depend on your project and budgets.

When setting up your cluster on EKS here are some links to the EKS documentation that can help you decide what type of setup to implement.

• Choosing an AWS container service
• Getting started with AWS Batch on Amazon EKS
• Amazon EKS service quotas
• Choosing an Amazon EC2 instance type
• Amazon EKS pricing

Installation

This step will install all the Kelvin services to your EKS cluster.

Setup EKS Cluster

To start you need to setup your own EKS cluster following Amazon's instructions. There are four basic steps to installing and setting up EKS clusters;

1. Create an EKS Cluster: Use the AWS Management Console, AWS CLI, or an AWS SDK to create an EKS cluster.
2. Configure kubectl: Set up kubectl, the Kubernetes command-line tool, with credentials to access your EKS cluster.
3. Create Worker Nodes: Launch and configure worker nodes that will join your cluster. This can be done through AWS Management Console or using AWS CloudFormation templates.
4. Connect Worker Nodes to Cluster: After creating the worker nodes, connect them to your EKS cluster.

Kuberenetes documentation website has a getting started guide for achieving all these steps; https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/

Install Cluster

Kelvin UIKelvin APIKelvin API Client (Python)

Let's dive in and show you how fast and easy the setup can be by first logging into the Kelvin Platform and going step by through the process.

1. Register Cluster in the Cloud

To start, go to the Orchestration page and click on the Register Cluster button;

Then in the popup you will see three cluster options to choose from.

Note

Basic users who do not want to maintain their own Kubernetes clusters can choose K3S. In this case Kelvin will install and manage all aspects of the cluster.

Some versions of Kelvin only allow the basic Docker environment create option.

For this tutorial we will select the Kubernetes option.

Then click on the Next button.

In step two, assign the new cluster a recognizable display name. The name id will be automatically created from your display name. It can usually be left unchanged.

Note

The Name ID will be created automatically when you create a Display Name. Normally this default name is satisfactory but you do have the option to customize it if you want.

Option	Description
Display Name	This can be any characters and spaces that gives your cluster a memorable name for reference
Name ID	A unique lower-case alphanumeric name which uniquely identifies this cluster. This will be automatically filled in when you type the Display Name. Normally you do not need to change unless the Name ID clashes with another cluster's Name ID.

The Name ID is the unique identifier name for the cluster. This must contain only lowercase alphanumeric characters. The ., _ and - characters are also allowed to separate words instead of a space BUT can not be at the beginning or end of the name.

Click Register.

You can then download the manifest YAML file and then run the CLI command in your EKS environment.

Make sure the manifest YAML file is locally accessible to the kubectl command.

Also you can see the sample kubectl command you will need to execute for your Cluster.

kubectl apply --server-side -f kelvin.yaml

Next click the Close button. The cluster is now registered in the Cloud and you can see it is pending registration.

2. Install Kelvin specific pods on your Cluster

Run the command to execute the yaml file parameters into your Cluster.

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.

kubectl apply --server-side -f kelvin.yaml

curl -X "POST" \
"https://<url.kelvin.ai>/api/v4/orchestration/clusters/create" \
-H "Authorization: Bearer <Your Current Token>" \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-d '{
    "name": "doc-demo-cluster-eks",
    "title": "Doc Demo Cluster EKS",
    "type": "kubernetes"
}'

The response will look something like this;

{
   "name":"doc-demo-cluster-eks",
   "title":"Doc Demo Cluster EKS",
   "type":"kubernetes",
   "ready":false,
   "status":"pending_provision",
   "last_seen":null,
   "sync_scrape_interval":120,
   "manifests_scrape_interval":130,
   "manifests_scrape_enabled":false,
   "version":{
      "k8s_version":"",
      "kelvin_version":""
   },
   "created":"2025-01-23T14:20:59.62627Z",
   "updated":"2025-01-23T14:20:59.62627Z",
   "service_account_token":"bm9kZS1jbGllbnQtZG9jLWRlbW8tY2x1c3Rlci1la3M6VE9VMWZsSmEwRUtNODZSSWM0RnI1ZFBXRDNYWjlnaUg=",
   "provision_script":"kubectl apply --server-side -f kelvin.yaml",
   "join_script":"",
   "telemetry_scrape_interval":110,
   "telemetry_enabled":false,
   "telemetry_buffer_size":10,
   "forward_logs_enabled":false,
   "forward_logs_buffer_size":10,
   "upgrade_status":{
      "state":"idle",
      "message":""
   },
   "upgrade_pre_download":false,
   "upgrade_instantly_apply":true
}

You can then get the manifest YAML file that is mentioned in the provision script with the API request;

curl -X "GET" \
"https://<url.kelvin.ai>/api/v4/orchestration/clusters/doc-demo-cluster-eks/provision/get" \
-H "Authorization: Bearer <Your Current Token>" \
-H "Accept: application/json" 

which will give a YAML response starting like this;

kind: Namespace
apiVersion: v1
metadata:
  name: kelvin-admin
---
apiVersion: v1
kind: Secret
type: kubernetes.io/tls
metadata:
  name: kelvin-certificate
  namespace: kelvin-admin
  labels:
    app.kubernetes.io/part-of: kelvin
data:
  ca.crt:

  ...

Copy and paste this into a file called kelvin.yaml and then run the CLI command given in the key provision_script in your EKS environment.

Make sure the manifest YAML file is locally accessible to the kubectl command.

The actual CLI command is;

kubectl apply --server-side -f kelvin.yaml

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.

from kelvin.api.client import Client

# Login
client = Client(url="https://<url.kelvin.ai>", username="<your_username>")
client.login(password="<your_password>")

# Set or Update the Default Cluster in an Instance
response = client.orchestration.create_orchestration_clusters(data={
    "name": "doc-demo-cluster-eks",
    "title": "Doc Demo Cluster EKS",
    "type": "kubernetes"
    })

print(response)

You can save the manifest YAML file that is mentioned in the provision script with;

kelvin_yaml = client.orchestration.get_orchestration_clusters_provision(cluster_name=response.name)
with open("kelvin.yaml", 'w') as file:
    file.write(str(kelvin_yaml))

Make sure the manifest YAML file is locally accessible to the kubectl command.

And then run the CLI command given in the key provision_script in your EKS environment. The actual CLI command is;

kubectl apply --server-side -f kelvin.yaml

Congratulations, after a few minutes depending on your Internet speed your new cluster is ready for use.