Cluster Autoscaler

 

AutoScaler

• Responsible for dynamically scale the nodes in the node group
• Runs as deployment
• Decides on the basis of CPU & RAM availability
• a mixture of on-demand and spot instance possible

LAB

The cluster autoscaler automatically launches additional worker nodes if more resources are needed, and shutdown worker nodes if they are underutilized. The autoscaling works within a nodegroup, hence create a nodegroup first which has this feature enabled.

Create AutoScale node group in AZ by creating below yaml file (eks-auto.yaml)

    apiVersion: eksctl.io/v1alpha5
    kind: ClusterConfig

    metadata:
      name: EKS-raman-cluster
      region: us-east-1

    nodeGroups:
      - name: scale-east1c
        instanceType: t2.small
        desiredCapacity: 1
        maxSize: 5
        availabilityZones: ["us-east-1c"]
        iam:
          withAddonPolicies:
            autoScaler: true
        labels:
          nodegroup-type: stateful-east1c
          instance-type: onDemand
        ssh: # use existing EC2 key
          publicKeyName: vdevops
      - name: scale-east1d
        instanceType: t2.small
        desiredCapacity: 1
        maxSize: 5
        availabilityZones: ["us-east-1a"]
        iam:
          withAddonPolicies:
            autoScaler: true
        labels:
          nodegroup-type: stateful-east1d
          instance-type: onDemand
        ssh: # use existing EC2 key
          publicKeyName: vdevops
      - name: scale-spot
        desiredCapacity: 1
        maxSize: 5
        instancesDistribution:
          instanceTypes: ["t2.small", "t3.small"]
          onDemandBaseCapacity: 0
          onDemandPercentageAboveBaseCapacity: 0
        availabilityZones: ["us-east-1c", "us-east-1a"]
        iam:
          withAddonPolicies:
            autoScaler: true
        labels:
          nodegroup-type: stateless-workload
          instance-type: spot
        ssh:
          publicKeyName: vdevops

Run Below command to create nodegroup with autoscaler enabled

    eksctl create nodegroup --config-file=eks-auto.yaml
    # Run below command to check the nodes get created or not
    kubectl get nodes
    # Run below command to delete ng-1 node group
    eksctl delete nodegroup -f eks.yaml --include="ng-1"
    eksctl delete nodegroup -f eks.yaml --include="ng-1" --approve

Deploy the AutoScaler

    kubectl apply -f https://raw.githubusercontent.com/kubernetes/autoscaler/master/cluster-autoscaler/cloudprovider/aws/examples/cluster-autoscaler-autodiscover.yaml

put required annotation to the deployment:

    kubectl -n kube-system annotate deployment.apps/cluster-autoscaler cluster-autoscaler.kubernetes.io/safe-to-evict="false"

get the autoscaler image version:

      open https://github.com/kubernetes/autoscaler/releases and get the latest release version matching your Kubernetes version, e.g. Kubernetes 1.14 => check for 1.14.n  where "n" is the latest release version. change cluster name (EKS-raman-cluster) and version of EKS
      
      kubectl -n kube-system edit deployment.apps/cluster-autoscaler

Create nginx-deployment.yaml

    apiVersion: apps/v1

    kind: Deployment

    metadata:

      name: test-autoscaler

    spec:

      selector:

        matchLabels:

          app: nginx

      replicas: 1

      template:

        metadata:

          labels:

            service: nginx

            app: nginx

        spec:

          containers:

          - image: nginx

            name: test-autoscaler

            resources:

              limits:

                cpu: 300m

                memory: 512Mi

              requests:

                cpu: 300m

                memory: 512Mi

          nodeSelector:

            instance-type: spot

Check the Log details

  kubectl -n kube-system logs deployment.apps/cluster-autoscaler