Skip to main content

Quick Start

This guide walks you through creating a Private Cluster and deploying your first workload on it.

Step 1: Navigate to Private Cluster​

  1. Go to the TIR Dashboard.
  2. In the left sidebar, click Private Cluster under Products.

Step 2: Create a Private Cluster​

  1. Click Create Private Cluster.
  2. Select the required compute configuration (GPU type and node count).
  3. Choose a pricing option:
    • Hourly – Flexible, pay-as-you-go pricing
    • Committed – Lower cost for predictable workloads with a fixed commitment period
  4. Enter the Cluster Name.
  5. Specify the Node Count.
  6. Click Create.

You can also create a Private Cluster using the API. Refer to the Private Cluster Creation API documentation for detailed request parameters, authentication requirements, and sample request/response examples.

Your cluster will be provisioned within a few minutes.

Step 3: Allocate Nodes to a Project​

Once the cluster is running, assign nodes to a project so your team can launch workloads.

  1. Open your cluster and navigate to Project Allocation.
  2. Locate the project you want to assign nodes to.
  3. Open the Actions menu and select Add Nodes.
  4. Specify the node count.
  5. Confirm the allocation.

Nodes move from the Free state to Allocated once assigned to a project.

Step 4: Launch a Workload​

With nodes allocated, your team can now launch services within the cluster:

  • Nodes – GPU-backed Jupyter notebook environments
  • Inference Endpoints – Production model serving
  • Training Clusters – Distributed training jobs
  • Vector Databases – Similarity search for RAG workflows
No additional charges are incurred for services deployed inside a Private Cluster.

Commitment Options (Committed Plans)​

If you selected a committed plan during creation, choose what happens when the commitment period ends:

OptionBehaviour
Auto-RenewAutomatically renews for another commitment period
Switch to Hourly BillingRetains the cluster at hourly rates after the period
Auto-TerminateAutomatically deletes the cluster when the period ends
tip

Committed plans provide cost benefits for steady workloads. Any additional nodes added after creation are billed at hourly rates.

Next Steps​