Deploy on AWS using Helm Charts

Step 0: IAM Role Setup

All three IAM roles must be created before the EKS cluster is provisioned.

0.1 Create Cluster IAM Role

Navigate to IAM Console > Roles > Create Role:

• Trusted Entity: AWS Service

• Use case: EKS — Cluster

Policies to Attach

Policy Name	Purpose
AmazonEKSClusterPolicy	Core EKS cluster management — grants the control plane permission to manage AWS resources on your behalf
AmazonEKSVPCResourceController	Required for Security Groups for Pods; allows cluster to manage ENIs and IPs
AmazonEC2ContainerRegistryReadOnly	Allows the cluster to pull container images from ECR
AmazonEKS_CNI_Policy	Enables VPC CNI networking for pod IP allocation
ElasticLoadBalancingFullAccess	Required only if the AWS Load Balancer Controller add-on is installed

---

0.2 Create Node IAM Role

Navigate to IAM Console > Roles > Create Role:

• Trusted Entity: AWS Service

• Use case: EC2

Policies to Attach

Policy Name	Purpose
AmazonEKSWorkerNodePolicy	Allows EC2 nodes to register with the EKS cluster and perform node-level operations
AmazonEKS_CNI_Policy	Enables the VPC CNI plugin to configure pod networking on nodes
AmazonEC2ContainerRegistryReadOnly	Allows nodes to pull container images from ECR

---

0.3 Create EBS CSI Driver IAM Role (Pod Identity Role)

Navigate to IAM Console > Roles > Create Role:

• Trusted Entity Type: Custom trust policy:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Principal": {
        "Service": "pods.eks.amazonaws.com"
      },
      "Action": [
        "sts:AssumeRole",
        "sts:TagSession"
      ]
    }
  ]
}

• Policies to Attach: AmazonEBSCSIDriverPolicy

• Role Name: AmazonEKSPodIdentityAmazonEBSCSIDriverRole

---

Step 1: Create an EKS Cluster

1.1 Create VPC Subnets (if not using an existing VPC)

Go to VPC Dashboard > Subnets > Create subnet:

• VPC: Choose your existing VPC (or create a new one first under VPC > Create VPC)

• Subnets: Create at least 2 subnets in different Availability Zones (AZs).

• CIDR Block Example:

• Subnet 1 (us-east-1a): 10.0.1.0/24

• Subnet 2 (us-east-1b): 10.0.2.0/24

• These CIDR blocks allow 256 IP addresses each, which is sufficient for many small to medium workloads.

• Ensure each subnet has Auto-assign public IPv4 address enabled (under subnet settings).

Tip: Choose AZs from the same region (e.g., us-east-1a, us-east-1b). This helps with high availability and proper node distribution.

1.2 Create EKS Cluster

Navigate to EKS > Clusters > Add cluster > Create:

• Cluster Name: (e.g., hypers-cluster)

• Disable EKS Auto Mode

• Cluster Authentication Mode: EKS API and ConfigMap (Recommended)

• Cluster Endpoint Access: Public and Private

1.3 Add Cluster Add-ons (during creation):

Enable the following:

• CoreDNS

• Kube-proxy

• Amazon VPC CNI

• Amazon EKS Pod Identity Agent

• Amazon EBS CSI Driver

• When selecting this, choose the IAM role AmazonEKSPodIdentityAmazonEBSCSIDriverRole created earlier

• External DNS (if required)

---

Step 2: Create and Configure Node Group

2.1 Navigate to Cluster > Compute > Add Node Group

• Name: hypers-node-group

• Select the Node IAM Role created earlier

2.2 Configure Node Group:

• Minimum Nodes: 4 (important to avoid Helm deadlocks)

• Instance Types: Select appropriate instance (e.g., t3.medium or higher)

• Subnets: Select the 2 subnets across different AZs created earlier

Wait for the node group to be fully created and active before proceeding.

---

Step 3: Configure CLI Environment

3.1 Set AWS Credentials

export AWS_ACCESS_KEY_ID=<Your_Access_Key_ID>
export AWS_SECRET_ACCESS_KEY=<Your_Secret_Access_Key>
export AWS_SESSION_TOKEN="<Your_AWS_SESSION_TOKEN>" # If using temporary credentials

3.2 Connect kubectl to EKS Cluster

aws eks update-kubeconfig --name <cluster_name> --region <region_name>

3.3 Apply Default Storage Class

Create a file storageclass.yaml with the following:

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: gp3
  annotations:
    storageclass.kubernetes.io/is-default-class: "true"
provisioner: ebs.csi.aws.com
parameters:
  type: gp3
reclaimPolicy: Delete
allowVolumeExpansion: true
volumeBindingMode: WaitForFirstConsumer

Apply the file:

kubectl apply -f storageclass.yaml

---

Step 4: Install Hyperswitch via Helm

4.1 Add Hyperswitch Helm Repo

helm repo add hyperswitch  https://juspay.github.io/hyperswitch-helm
helm repo update 

4.2 Install the Hyperswitch Stack

helm install hypers-v1 hyperswitch/hyperswitch-stack -n hyperswitch --create-namespace

---

Step 5: Access Hyperswitch Services

Use the commands from the Step 4.2 (from above) for port-forwarding to access the services.

Step 6: Cleanup and Uninstallation

To uninstall Hyperswitch and remove the namespace:

helm uninstall hypers-v1 -n hyperswitch

kubectl delete namespace hyperswitch

---

Notes

• Ensure all IAM roles are correctly attached and selected during EKS cluster and addon setup.

• Having fewer than 4 nodes may cause timeout or deployment issues with Helm.

• You can view logs, pods, and service states using kubectl get pods -n hyperswitch, kubectl logs, and similar commands for debugging.

---

You are now ready to deploy and manage the Hyperswitch stack on AWS EKS using Helm via the AWS Console!