Deploying Spotfire on Amazon Elastic Kubernetes Service (EKS)

This page provides an overview of the main steps needed to prepare an environment in AWS and to deploy the Spotfire Platform ⧉ on Amazon Elastic Kubernetes Service (EKS) ⧉, using the Spotfire CDK ⧉.

You will deploy the Spotfire Platform on AWS using the following services:

• Kubernetes cluster: Amazon Elastic Kubernetes Service (EKS) ⧉.
• Database server: Amazon RDS for PostgreSQL ⧉.
• Blob storage: Amazon S3 ⧉.
• Load balancer: Amazon Elastic Load Balancer (ELB) ⧉.

This is a quickstart guide. For more information, see the official documentation. Always follow the documentation and recommended best practices from the vendor.

Remember to change the provided example values to adapt them to your own environment and needs.

Prerequisites

• An account in AWS with permissions for the required services
• A Linux host with the following clients installed:
  • Amazon CLI ⧉
  • AWS ekscli ⧉
  • Kubectl ⧉
  • Helm 3+ ⧉

Note: The AWS CLI (aws) and ekscli are used in the examples below, but you can use the AWS web interface, REST API, libraries or any other available methods.

Steps

1. Prepare your AWS environment

• Login to your AWS account and set up your project:

  aws sso login
  export REGION=eu-north-1
  

2. Create an EKS cluster

1. Define the variables for your cluster:

   export CLUSTER_NAME=my-eks
   export NODE_TYPE=m5.xlarge
   export NUM_NODES=2
   

   Note: In this example we create a 3-nodes cluster using the m5.xlarge virtual machine type (8 vCPUs, 32 GB). See the Amazon EC2 Instance types ⧉ to understand which SKU you need for your K8s nodes. The AWS M-Family of machines are general purpose VMs for Enterprise-grade containerized applications.

   Note: See the Spotfire system requirements ⧉ for the minimum and recommended sizing. Observe your K8s resource utilization to understand which node size and how many nodes do you need.

2. Create the Amazon Elastic Kubernetes Service (EKS):

   eksctl create cluster \
       --name $CLUSTER_NAME \
       --region $REGION \
       --nodegroup-name $CLUSTER_NAME-nodes \
       --node-type $NODE_TYPE \
       --nodes $NUM_NODES
   

   It will take ~10-15 minutes to create the K8s cluster.

   For more information, see Get started with Amazon EKS – eksctl ⧉.

3. After creating the cluster, you must configure kubectl to use the new EKS cluster:

   aws eks update-kubeconfig \
      --region $REGION \
      --name $CLUSTER_NAME
   

4. Verify that you can connect to the cluster using kubectl:

   kubectl get nodes -o wide
   aws eks describe-cluster \
       --region $REGION \
       --name $CLUSTER_NAME
   

For more information, see the Amazon Elastic Kubernetes Service Documentation ⧉.

You can learn about reference architectures, diagrams, and best practices in the AWS Architecture Center ⧉.

See also the Amazon EKS Best Practices Guide ⧉.

3. Deploy Spotfire

1. Create a namespace for your deployment:

   export NAMESPACE=spotfire-quickstart
   kubectl create namespace $NAMESPACE
   

2. Create a pull secret for the Spotfire container registry:

   export REGISTRY_SERVER=oci.spotfire.com
   export REGISTRY_SECRET=spotfire-oci-secret
   export REGISTRY_USERNAME=<username>
   export REGISTRY_PASSWORD=<password>
   
   kubectl create secret docker-registry $REGISTRY_SECRET \
       --namespace $NAMESPACE \
       --docker-server=$REGISTRY_SERVER \
       --docker-username=$REGISTRY_USERNAME \
       --docker-password=$REGISTRY_PASSWORD
   

   The secret is used by the Kubernetes cluster to pull the Spotfire container images from the Spotfire OCI registry.

3. Log in to the Spotfire Helm charts registry:

   helm registry login -u $REGISTRY_USERNAME oci.spotfire.com/charts
   

   This is needed to access the Spotfire Helm charts in the Spotfire OCI registry.

4. Deploy the Spotfire Platform using the spotfire-platform Helm chart. For example:

   export MY_SPOTFIRE_RELEASE=vanilla-spotfire
   
   helm upgrade --install $MY_SPOTFIRE_RELEASE \
       oci://$REGISTRY_SERVER/charts/spotfire-platform \
       --version 2.0.0 \
       --namespace=$NAMESPACE \
       --set global.spotfire.acceptEUA=true \
       --set global.spotfire.image.registry=$REGISTRY_SERVER \
       --set global.spotfire.image.pullSecrets[0]=$REGISTRY_SECRET \
       --set spotfire-server.configuration.site.publicAddress=http://spotfire.example.com \
       --set spotfire-server.postgresql.enabled=true \
       ...
   

   For more information, see the spotfire-platform Helm chart.

You have now deployed the Spotfire platform on AWS, using Amazon Elastic Kubernetes Service (EKS).

4. Next steps

You can now continue with:

• Configuring an Amazon Network Load Balancer or an Amazon Application Load Balancer.
• Configuring an Amazon RDS for PostgreSQL database as the Spotfire database.
• Configuring an Amazon S3 bucket as the Spotfire external library storage.

Cleanup

To avoid unneeded resource usage, once you have completed these tutorials, delete any created resources:

eksctl delete cluster --name $CLUSTER_NAME
...

For more information, see Delete an Amazon EKS cluster ⧉.