Cluster Actions

ScaleOps supports defining default policy and automation for the entire cluster using Helm chart values.

Configure the clusterAutomation section in your values.yaml to specify the desired state for cluster workloads.

Configuration Available in v1.27.1+


clusterAutomation:
  enabled: true
  rightsize:
    optimize: true               # Automate Workload Rightsizing
  spot:
    optimize: true               # Automate Spot Optimization
  replicas:
    optimize: true               # Automate Replicas Optimization
  podPlacement:
    optimize: true               # Automate Pod Placement
  java:
    observability: true          # Enable Java observability cluster-wide
    optimize: true               # Automate Java Optimization

Fields

`clusterAutomation.enabled`

Must be true to apply the Helm-defined cluster automation settings. Otherwise, these settings have no effect.

Automation Fields

Each product section supports an optimize field that enables automation for all workloads in the cluster:

clusterAutomation.rightsize.optimize: Enables Workload Rightsizing automation
clusterAutomation.spot.optimize: Enables Spot Optimization automation
clusterAutomation.replicas.optimize: Enables Replicas Optimization automation
clusterAutomation.podPlacement.optimize: Enables Pod Placement automation
clusterAutomation.java.observability: Enables Java observability cluster-wide
clusterAutomation.java.optimize: Enables Java Optimization automation

Default Policy Fields

Set default policies for different optimization features:

clusterAutomation.rightsize.defaultPolicy: Sets default policy for Workload Rightsizing
clusterAutomation.spot.defaultPolicy: Sets default policy for Spot Optimization
clusterAutomation.replicas.defaultPolicy: Sets default policy for Replicas Optimization

ℹ️

When any optimize field is set to true, ScaleOps will create an AutomatedNamespace CR for each namespace. Disabling automation does not revert or delete the AutomatedNamespace CR for each namespace.

ConfigMap Configuration (Deprecated)

You can also configure cluster automation using a ConfigMap. This method is maintained for backward compatibility.


apiVersion: v1
data:
  rightsize-optimize: "true"          # Automate Cluster Workload Rightsizing
  replicas-optimize: "true"           # Automate Cluster Replicas Optimization
  pod-placement-optimize: "true"      # Automate Cluster Pod Placement
  spot-optimize: "true"               # Automate Cluster Spot Optimization
  gpu-optimize: "true"                # Automate Cluster GPU Rightsizing
kind: ConfigMap
metadata:
  name: scaleops-cluster-operations
  namespace: scaleops-system

Default Policy Fields

The following fields set default policies for different optimization features:

default-rightsize-policy: Sets default policy for Workload Rightsizing
default-replicas-policy: Sets default policy for Replicas Optimization
default-spot-policy: Sets default policy for Spot Optimization
default-gpu-policy: Sets default policy for GPU Rightsizing

ℹ️

When any field is set to true within the ConfigMap, ScaleOps will create an AutomatedNamespace CR for each namespace. Deleting the ConfigMap or setting the field to false does not revert or delete the AutomatedNamespace CR for each namespace.

Backward Compatibility

The optimize field is maintained for backward compatibility and applies only to Workload Rightsizing. It is equivalent to setting rightsize-optimize: "true".
The defaultPolicy field is maintained for backward compatibility and sets default policy only to Workload Rightsizing. It is equivalent to setting default-rightsize-policy.