Skip to main content

Cluster Management

Manage your private cluster and sync resources between the public cloud and your cluster.

Service Activation Required

This is a value-added feature that needs to be paid for additionally. Contact us to enable access.

Typical Workflow

  1. Register Cluster - Add and configure your Kubernetes cluster
  2. View Node Details - Check available nodes and resources
  3. Create Sync Task - Sync models or datasets to your cluster

cluster-entry


Register Cluster

Add your Kubernetes cluster to Smart Studio.

  1. Click Register Cluster.
  2. Enter your Cluster Name and paste your KubeConfig.
  3. Click Test Connection to validate the configuration.
  4. Click Register Cluster to complete the setup.

config-cluster


After registering, select a cluster from the list to view details or create sync tasks.

View Cluster Details

View your cluster information, node list, and resource utilization.


cluster-details


Node List

View all nodes in your Kubernetes cluster.

FieldDescription
NameThe name of the Kubernetes node
StatusNode status (Ready, NotReady, Unknown)
GPU ResourceGPU resource type and capacity
Allocated GPUsNumber of GPUs allocated to workloads
Allocated CPUsNumber of CPUs allocated to workloads
GPU UtilizationCurrent GPU utilization percentage
GPU MemoryGPU memory usage

Network Accessibility

View the Internal and External API Server Endpoint information for your cluster.

Real-time Resource Utilization

Monitor your cluster's resource usage in real-time:

GPU Utilization, GPU Memory, CPU Utilization, CPU Memory

Create Sync Task

Create sync tasks to transfer models between the public cloud and your cluster.


cluster-sync-list


Sync Task Configuration

  • Sync Type: Sync to the Cluster
  • Purpose: For deployment or For fine-tuning
  • Source Model: Select the model to sync from the public cloud

create-sync-task


Models Sync List

After creating sync tasks, track their progress in the Models tab:

ColumnDescription
Model NameThe name of the synced model
SIZETotal size of the model files
STORAGE PATHThe storage location on your cluster
Sync StatusCurrent sync progress (Syncing, Completed, Failed)
Last SyncTimestamp of the last sync operation

Datasets Sync List

Track dataset sync tasks in the Datasets tab:

ColumnDescription
Dataset NameThe name of the synced dataset
SizeTotal size of the dataset
Sync StatusCurrent sync progress
Last SyncTimestamp of the last sync operation

Node Details

Select a node from the Node List to view detailed resource information.

GPU Details

View the GPU processes running on a specific node:

ColumnDescription
Pod NameThe Kubernetes pod using the GPU
GPU UtilizationPer-pod GPU utilization percentage
GPU MemoryPer-pod GPU memory usage

Next Steps

Configure Secrets

Add your API key in Secrets to create a unified API for accessing all Model resources on the platform.

Try Model Lab

Find the perfect model for your use case by testing different configurations in our interactive playground.