Skip to main content

Kubernetes Scale

Last updated on

The Kubernetes Scale step changes the number of running pods for a workload to a target count or percentage. Use it to scale up before a load event, scale down to save resources, or adjust a canary workload during a canary stage.

Before you begin​

Before you configure the step, make sure you have the following in place:

  • A Kubernetes service: Go to Kubernetes services to set up service manifests and an artifact source.
  • A Kubernetes infrastructure: Go to Kubernetes infrastructure to connect your cluster and namespace.
  • A Harness delegate in your target cluster: The delegate runs deployment steps in the cluster.
  • Runtime configuration: Every Kubernetes stage requires a runtime block specifying a connector and namespace. Go to Kubernetes runtime configuration to understand the required fields.

Add the Kubernetes Scale step​

To add the step:

  1. In your pipeline, go to the Kubernetes stage.
  2. Select + Add Step in the execution section.
  3. Search for Kubernetes Scale and select it.
  4. Configure the step parameters described below.
  5. Select Apply Changes.

Configure the step​

The following parameters are available on the Kubernetes Scale step.

ParameterDescriptionRequired
NameDisplay name for the step in the pipeline.Required
WorkloadThe workload to scale, in [namespace/]Kind/Name format. For example, default/Deployment/my-app.Required
Instance SelectionWhether to scale by instance count or percentage. Select Count or Percentage.Required
InstancesThe target number of pods (when Count is selected) or the percentage of current replicas to scale to (when Percentage is selected).Required
Kubeconfig PathPath to the kubeconfig file, derived from the infrastructure configuration. Default: ${{infra.kube_config_path}}.Optional
NamespaceDefault namespace to use when the workload field does not include one.Optional
Release NameRelease name for pod label lookup. Default: ${{infra.releaseName}}.Optional
TimeoutMaximum time the step can run before it is marked as failed. Default: 5m.Optional

Set the workload​

Enter the workload in [namespace/]Kind/Name format:

  • default/Deployment/my-app: scales the Deployment named my-app in the default namespace
  • Deployment/my-app: uses the namespace configured in the Namespace field or derived from the infrastructure

Supported workload types are Deployment and DaemonSet. Only one workload can be specified per step.

You can use a Harness expression in the Workload field to reference a workload from a preceding step. This is useful in canary deployments where you target the canary workload by name:

<+stages.[Stage_Id].spec.execution.steps.[Step_Id].output.outputVariables.canaryWorkload>

Configure instance count or percentage​

Count scales the workload to the exact number of pods you enter.

Percentage scales to a percentage of the workload's current replica count. For example, if the workload has 10 replicas and you enter 50, the step scales it to 5 replicas.

Percentage must be a whole number

Harness does not support decimal percentages. Enter whole numbers only, for example, 50, not 50.5. The step fails if a decimal value is provided.

To remove all running pods without deleting the workload resource, enter 0 in the Instances field.

YAML example​

- name: Kubernetes Scale
  id: k8sScaleStep
  template:
    uses: k8sScaleStep
    with:
      workload: "default/Deployment/my-app"
      instances: "2"
      instances_unit_type: "count"

To scale to a percentage of current replicas:

- name: Kubernetes Scale
  id: k8sScaleStep
  template:
    uses: k8sScaleStep
    with:
      workload: "default/Deployment/my-app"
      instances: "50"
      instances_unit_type: "percentage"

Read the step output​

The step log shows the current replica count, the target replica count, and the result of the scale operation.

After the step runs, reference the pod list in downstream steps:

<+steps.[Step_Id].output.outputVariables.pods>

Advanced settings​

The following advanced settings are available on the Kubernetes Scale step.

  • Timeout duration: Maximum time the step is allowed to run before being terminated.
  • On failure: Define what happens if the step fails, such as retry, mark as success, or abort.
  • Strategy: Configure a looping strategy to run this step over a list of values.
  • Conditional execution: Run this step only when a specified condition is true.

Limitations in the unified platform​

Unsupported features in the unified platform

The following feature is available in the standard Harness Kubernetes Scale step but is not supported in the unified platform.

Skip Steady State Check: In the standard platform, the Scale step includes a built-in steady-state check that runs after scaling, controlled by the skipSteadyStateCheck flag. In the unified platform, this check is not built into the Scale step. To verify workload health after scaling, add a Kubernetes Steady State Check step after the Scale step.

Next steps​

Last updated on