Create and Manage Compute Configs

This guide describes how to create and manage Compute Configs. A Compute Config is a reusable blueprint of cluster settings that PaletteAI applies when provisioning Compute Pools, so users do not re-enter the same node and networking values for every deployment.

You can create Compute Configs using the PaletteAI UI or using YAML Kubernetes manifests.

Prerequisites

UI Workflow

• A Project or Tenant in Ready status.

• The spectrocloud.com/computeconfigs:create permission. If the Create Compute Config button is not displayed, your role does not include the required permission.

YAML Workflow

• A Project or Tenant in Ready status.

• Access to the hub cluster with permissions to create ComputeConfig resources in the target namespace.

Create a Compute Config

UI Workflow

1. Log in to the PaletteAI console. Ensure you are in the correct Project or Tenant scope.

2. From the left main menu, select Settings, and then select the Compute Config tab.

3. Select Create Compute Config.

4. On the Basic Information page, enter a Compute Config Name. At the tenant scope, you can also choose which Projects the Compute Config is shared with. Select Next.

5. (Optional) On the Edge Configuration page, configure cluster-wide Edge defaults such as NTP servers and network overlay settings. Select Next.

6. On the Node Pool Defaults page, configure the default control plane and worker pool settings.

   Control Plane Pool:

   Field	Description	Required
   Node Count	Number of control plane nodes. Valid values: 1, 3, or 5.	✅
   Single node cluster	Toggle to deploy clusters as single-node, where one node serves as both control plane and worker.	❌
   Run workloads on control plane	Toggle to allow workloads on control plane nodes. Required for single-node clusters.	❌
   Architecture	CPU architecture. AMD64 (default) or ARM64.	❌
   CPU Count	Per-node minimum CPU count.	❌
   Memory	Per-node minimum memory.	❌
   Annotations / Labels	Expand Metadata to add key-value pairs to control plane nodes.	❌
   Taints	Expand to add node taints. Each taint requires a key, value, and effect (NoSchedule, PreferNoSchedule, or NoExecute).	❌

   Worker Pools:

   Select Add Worker Pool to add default worker pools. For each pool, configure the Architecture, Min Worker Nodes, CPU Count, Memory, GPU requirements (GPU Variant, GPU Count, GPU Memory), and optional Metadata and Taints.

   info

   Control plane taints and worker pool taints are mutually exclusive. If the control plane defines taints, worker pools cannot.

   Select Next.

7. On the Review page, confirm your settings, and then submit the Compute Config.

YAML Workflow

A ComputeConfig is a namespaced resource in the spectrocloud.com/v1alpha1 API group. Only the control plane pool's nodeCount is required; everything else refines the defaults.

1. Create a ComputeConfig manifest.

   compute-config.yaml

   apiVersion: spectrocloud.com/v1alpha1
   kind: ComputeConfig
   metadata:
     name: <compute-config-name>
     namespace: <project-namespace>
   spec:
     paletteClusterConfig:
       defaults:
         controlPlanePool:
           nodeCount: 3
           architecture: AMD64
           cpu:
             cpuCount: 4
           memory:
             memory: '16 GB'
         workerPools:
           - name: default-workers
             architecture: AMD64
             minWorkerNodes: 1

   The manifest breaks down as follows:

   • defaults.controlPlanePool sets the control plane defaults. nodeCount must be 1, 3, or 5.

   • defaults.workerPools lists default worker pool shapes. Add GPU requirements (gpu.variant, gpu.gpuCount) for GPU pools.

   • An optional edge block holds cluster-wide Edge defaults such as ntpServers and networkOverlayConfig, and defaults can also carry sshKeys and a deletionPolicy.

   For all fields, complete examples, and validation rules, refer to the ComputeConfig Configuration reference.

2. Apply the manifest to the hub cluster.

   kubectl apply --filename compute-config.yaml

Validate

UI Workflow

1. From the left main menu, select Settings, and then select the Compute Config tab.

2. Confirm the new Compute Config appears in the table.

3. To confirm it is usable, start the Compute Pool creation workflow: on the Node config page, select the gear icon to open the Advanced settings drawer and verify the Compute Config appears in the Compute Config drop-down menu.

YAML Workflow

1. Confirm the ComputeConfig resource exists.

   kubectl get computeconfig <compute-config-name> --namespace <project-namespace>

   Example Output

   NAME                   AGE
   my-compute-config      10s

Update or Delete a Compute Config

UI Workflow

1. From the left main menu, select Settings, and then select the Compute Config tab.

2. Beside the Compute Config, select the three-dot menu, and choose Edit or Delete.

Changes apply to Compute Pools created after the update; existing Compute Pools keep the values they were provisioned with.

YAML Workflow

Update the manifest and re-apply it, or delete the resource:

kubectl apply --filename compute-config.yaml

kubectl delete computeconfig <compute-config-name> --namespace <project-namespace>

Next Steps

• Create and Manage Compute Pools — Apply your Compute Config when provisioning a pool.

• ComputeConfig Configuration — All fields, examples, and validation rules.

• Deploy a Single-Node Compute Pool — A Compute Config recipe for single-node clusters.