Deploy a sample inflate workload to an Amazon EKS Auto Mode cluster

In this tutorial, you’ll learn how to deploy a sample workload to an EKS Auto Mode cluster and observe how it automatically provisions the required compute resources. You’ll use kubectl commands to watch the cluster’s behavior and see firsthand how Auto Mode simplifies Kubernetes operations on AWS. By the end of this tutorial, you’ll understand how EKS Auto Mode responds to workload deployments by automatically managing the underlying compute resources, without requiring manual node group configuration.

Prerequisites

Step 1: Review existing compute resources (optional)

First, use kubectl to list the node pools on your cluster.

kubectl get nodepools

Sample Output:

general-purpose

In this tutorial, we will deploy a workload configured to use the general-purpose node pool. This node pool is built into EKS Auto Mode, and includes reasonable defaults for general workloads, such as microservices and web apps. You can create your own node pool. For more information, see Create a Node Pool for EKS Auto Mode.

Second, use kubectl to list the nodes connected to your cluster.

kubectl get nodes

If you just created an EKS Auto Mode cluster, you will have no nodes.

In this tutorial you will deploy a sample workload. If you have no nodes, or the workload cannot fit on existing nodes, EKS Auto Mode will provision a new node.

Step 2: Deploy a sample application to the cluster

Review the following Kubernetes Deployment and save it as inflate.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: inflate
spec:
  replicas: 1
  selector:
    matchLabels:
      app: inflate
  template:
    metadata:
      labels:
        app: inflate
    spec:
      terminationGracePeriodSeconds: 0
      nodeSelector:
        eks.amazonaws.com/compute-type: auto
      securityContext:
        runAsUser: 1000
        runAsGroup: 3000
        fsGroup: 2000
      containers:
        - name: inflate
          image: public.ecr.aws/eks-distro/kubernetes/pause:3.7
          resources:
            requests:
              cpu: 1
          securityContext:
            allowPrivilegeEscalation: false

Note the eks.amazonaws.com/compute-type: auto selector requires the workload be deployed on an Amazon EKS Auto Mode node.

Apply the Deployment to your cluster.

kubectl apply -f inflate.yaml

Step 3: Watch Kubernetes Events

Use the following command to watch Kubernetes events, including creating a new node. Use ctrl+c to stop watching events.

kubectl get events -w --sort-by '.lastTimestamp'

Use kubectl to list the nodes connected to your cluster again. Note the newly created node.

kubectl get nodes

Step 4: View nodes and instances in the AWS console

You can view EKS Auto Mode Nodes in the EKS console, and the associated EC2 instances in the EC2 console.

EC2 Instances deployed by EKS Auto Mode are restricted. You cannot run arbitrary commands on EKS Auto Mode nodes.

Step 5: Delete the deployment

Use kubectl to delete the sample deployment

kubectl delete -f deployment.yaml

If you have no other workloads deployed to your cluster, the node created by EKS Auto Mode will be empty.

In the default configration, EKS Auto Mode detects nodes that have been empty for thirty seconds, and terminates them.

Use kubectl or the EC2 console to confirm the associated instance has been deleted.