Managing autoscale policies

Autoscale policies automatically expand or shrink a cluster in your private cloud based on factors like CPU, memory, and storage utilization thresholds. According to the applied autoscale policy for a cluster, Google Cloud VMware Engine monitors the cluster based on defined metrics and adds or removes nodes from the cluster automatically. Autoscale policies are not available on clusters containing a single node.

Autoscaling stretched private clouds

You can configure autoscaling policies on a stretched private cluster. The default autoscaling policies available for standard clusters are also available for stretched clusters. The same number of nodes (as specified in the autoscale policy) are added to or removed from each data zone of a stretched cluster to keep the stretched cluster configuration symmetrical.

Predefined policies

For each cluster in your private cloud, you can choose to apply an autoscale policy from a list of predefined policies. When metrics specified in an applied autoscale policy are sustained for a period of 30 minutes, VMware Engine triggers an action based on the policy.

Because compute, memory, and storage usage are often independent, autoscale policies that monitor multiple metrics use OR logic for node addition and AND logic for node removal.

The following table describes predefined autoscale policies that you can apply to a cluster:

Autoscale policy Add node condition Remove node condition
Storage Capacity Optimization Policy Storage consumed > 80% Storage consumed < 40%
CPU Performance Optimization Policy CPU utilization > 70% OR
Storage consumed > 80% 		CPU utilization < 45% AND
Storage consumed < 40%
Memory Performance Optimization Policy Memory utilization > 80% OR
Storage consumed > 80% 		Memory utilization < 45% AND
Storage consumed < 40%
CPU and Memory Performance Optimization Policy CPU utilization > 70% OR
Memory utilization > 80% OR
Storage consumed > 80% 		CPU utilization < 45% AND
Memory utilization < 45% AND
Storage consumed < 40%

Updating an autoscale policy

To view or change the autoscale policy associated with a cluster, do the following:

  1. Access the VMware Engine portal.
  2. Go to Resources and select a private cloud.
  3. Go to Clusters and click the Edit icon.
  4. On the page that appears, you can view your current policy, select a new one from the Autoscale policy drop-down list, and optionally adjust policy details.
  5. Click Submit once you're finished.

The following list describes each field during policy creation in more detail:

  • Add nodes (standard private cluster): Condition that must be met for a certain period of time before VMware Engine adds one or more nodes to the cluster.

  • Add nodes (stretched private cluster): Condition that must be met for a certain period of time before VMware Engine adds two or more (must be an even number of) nodes to the cluster. An equal number of nodes are added to each data zone of a stretched cluster.

  • Remove nodes (standard private cluster): Condition that must be met for a certain period of time before VMware Engine removes a node from the cluster.

  • Remove nodes (stretched private cluster): Condition that must be met for a certain period of time before VMware Engine removes two nodes from the cluster. An equal number of nodes are removed from each data zone of a stretched cluster.

  • Cool off period: Minimum amount of time between actions taken due to an autoscale policy.

  • Node limits (standard private cluster): Minimum and maximum number of nodes in a cluster between which an autoscale policy can adjust. The minimum number of data nodes in a standard private cluster is three (3).

  • Node limits (stretched private cluster): Minimum and maximum number of nodes in a cluster between which an autoscale policy can adjust.The minimum number of data nodes in a stretched cluster is six (6). The maximum data nodes in a stretched cluster is 32.