How to apply Node Labels and Taints on VKS Workload Cluster Nodes (v1beta)

Products

• Tanzu Kubernetes Runtime

Issue/Introduction

VKS workload clusters allow for the configuration of node labels and taints. These configurations are used to categorize workloads, manage node behavior, and control scheduling.

The node variable, which can be defined at the cluster level, offers flexibility. It can be overridden at the control plane or on a per-node pool basis, enabling the definition of specific values for particular node groups.

Resolution

Node labels and taints are defined using the node variable in the cluster specification.

Example configuration:

variables:
- name: node
  value:
    labels:
      tenant: tenant-foo
      organization: engineering
      managed: ""
    taints:
    - key: key1
      value: value1
      effect: NoSchedule
    - key: key2
      value: value2
      effect: NoExecute

or

workers:
      machineDeployments:
      - class: node-pool
        name: nodepool-test
        replicas: 1
        variables:
          overrides:
          - name: node
            value:
              labels:
                tenant: tenant-foo

When configured at the cluster level, the labels and taints apply to all nodes within the cluster.

When configured at the control plane or node pool level, they override the cluster-level settings for that specific node group.

Ensure the configuration follows the correct YAML syntax and v1beta API structure for successful application.