Configure Cluster Autoscaler (CA)

Cluster Autoscaler for AWS provides integration with Auto Scaling groups. It enables users to choose from four different options of deployment:

  • One Auto Scaling group - This is what we will use
  • Multiple Auto Scaling groups
  • Auto-Discovery
  • Master Node setup

Configure the Cluster Autoscaler (CA)

We have provided a manifest file to deploy the CA. Copy the commands below into your Cloud9 Terminal.

mkdir ~/environment/cluster-autoscaler
cd ~/environment/cluster-autoscaler
wget https://eksworkshop.com/scaling/deploy_ca.files/cluster_autoscaler.yml

Configure the ASG

We will need to provide the name of the Autoscaling Group that we want CA to manipulate. Collect the name of the Auto Scaling Group (ASG) containing your worker nodes. Record the name somewhere. We will use this later in the manifest file.

You can find it in the console by following this link.

ASG

Check the box beside the ASG and click Actions and Edit

Change the following settings:

  • Min: 2
  • Max: 8

ASG Config

Click Save

Configure the Cluster Autoscaler

Using the file browser on the left, open cluster_autoscaler.yml

Search for command: and within this block, replace the placeholder text <AUTOSCALING GROUP NAME> with the ASG name that you copied in the previous step. Also, update AWS_REGION value to reflect the region you are using and Save the file.

command:
  - ./cluster-autoscaler
  - --v=4
  - --stderrthreshold=info
  - --cloud-provider=aws
  - --skip-nodes-with-local-storage=false
  - --nodes=2:8:eksctl-eksworkshop-eksctl-nodegroup-0-NodeGroup-SQG8QDVSR73G
env:
  - name: AWS_REGION
    value: us-east-1

This command contains all of the configuration for the Cluster Autoscaler. The primary config is the --nodes flag. This specifies the minimum nodes (2), max nodes (8) and ASG Name.

Although Cluster Autoscaler is the de facto standard for automatic scaling in K8s, it is not part of the main release. We deploy it like any other pod in the kube-system namespace, similar to other management pods.

Create an IAM Policy

We need to configure an inline policy and add it to the EC2 instance profile of the worker nodes

Collect the Instance Profile and Role NAME from the CloudFormation Stack

INSTANCE_PROFILE_PREFIX=$(aws cloudformation describe-stacks | jq -r '.Stacks[].StackName' | grep eksctl-eksworkshop-eksctl-nodegroup)
INSTANCE_PROFILE_NAME=$(aws iam list-instance-profiles | jq -r '.InstanceProfiles[].InstanceProfileName' | grep $INSTANCE_PROFILE_PREFIX)
ROLE_NAME=$(aws iam get-instance-profile --instance-profile-name $INSTANCE_PROFILE_NAME | jq -r '.InstanceProfile.Roles[] | .RoleName')

mkdir ~/environment/asg_policy
cat <<EoF > ~/environment/asg_policy/k8s-asg-policy.json
{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Effect": "Allow",
      "Action": [
        "autoscaling:DescribeAutoScalingGroups",
        "autoscaling:DescribeAutoScalingInstances",
        "autoscaling:SetDesiredCapacity",
        "autoscaling:TerminateInstanceInAutoScalingGroup"
      ],
      "Resource": "*"
    }
  ]
}
EoF
aws iam put-role-policy --role-name $ROLE_NAME --policy-name ASG-Policy-For-Worker --policy-document file://~/environment/asg_policy/k8s-asg-policy.json

Validate that the policy is attached to the role

aws iam get-role-policy --role-name $ROLE_NAME --policy-name ASG-Policy-For-Worker

Deploy the Cluster Autoscaler

kubectl apply -f ~/environment/cluster-autoscaler/cluster_autoscaler.yml

Watch the logs

kubectl logs -f deployment/cluster-autoscaler -n kube-system

We are now ready to scale our cluster

  • cluster_autoscaler.yml (3 ko)