AWS Autoscaler Addon

Introduction

The Platform9 managed AWS Autoscaler addon object's primary function is to ensure the resources of the Kubernetes cluster are never exhausted. More specific information about Kubernetes Autoscaler can be found in its specific GitHub pages.

Description

The AWS autoscaler addon continually verifies the number of needed pods that are pending allocation. If there are pods that are pending allocation and the autoscaler addon is not at its defined capacity, new nodes are then provisioned to accommodate the demand. When demand drops and fewer nodes are needed, the autoscaler removes the unused nodes. After we install the autoscaler addon, its behavior is automatic, so users only need to define the number of replicas to be deployed.

Settings

The AWS Autoscaler settings are defined by information stored in a ConfigMap. When the addon is initially configured, user settings are provided by the client to support, who then implements the changes.

YAML Specifications

The optional resizer Addon was recently introduced in this version. If the enableAddonResizer field is not explicitly enabled, it will fall back to a single autoscaler container. The addon resizer has 6 configuration parameters, which will default to the values noted below if they are not provided.

apiVersion: sunpike.platform9.com/v1alpha2

kind: ClusterAddon

metadata:

  labels:

    sunpike.pf9.io/cluster: '<clsuuid>'

    type: cluster-auto-scaler-aws

  name: <clsuuid>-cluster-auto-scaler-aws

  namespace: default

spec:

  clusterID: '<clsuuid>'

  override:

    params:

    - name: clusterUUID

      value: ''

    - name: clusterRegion

      value: '<aws region>'

    # If enableAddonResizer is not set, the following values are explicitly used

    # If enableAddonResizer is set, the following 4 values are irrelevant and can be removed

    - name: cpuRequest

      value: 100m

    - name: cpuLimit

      value: 200m

    - name: memRequest

      value: 300Mi

    - name: memLimit

      value: 600Mi      

    # Addon resizer can be enabled by setting enableAddonResizer to any value

    # If enableAddonResizer is not set, the all following params are irrelevant

    # All of the following params are optional

    # cpu request/limit = .cpuRequestLimit + .cpuExtra * .minClusterSize

    # mem request/limit = .memRequestLimit + .memExtra * .minClusterSize

    # The values set here are the default values used if these params are not present    

    - name: enableAddonResizer

      value: "true"    

    - name: cpuRequestLimit

      value: "150m"

    - name: cpuExtra

      value: "0.5m"

    - name: memRequestLimit

      value: "300Mi"

    - name: memExtra

      value: "10Mi"

    - name: percentThreshold

      value: "5"      

    - name: minClusterSize

      value: "2"

  type: cluster-auto-scaler-aws

  version: 1.14.8

  watch: true

It should be noted that these addons are updated along with the pf9-kube version and can be referenced in the Support Matrix article.

Conclusion

Should you encounter any difficulties adopting or deploying the information noted above, our talented and responsive support team is available 24/7/365 to provide any additional assistance needed.