Skip to main content

Create a new Notebook Engine

This guide walks you through creating a new Notebook Engine in the H2O AI Cloud web interface.

What is a Notebook Engine?​

A Notebook Engine is a notebook authoring application and editing environment integrated with H2O products, including popular frameworks for machine learning, data science, and visualization. It provides:

  • Interactive notebook environment for data science workflows
  • Integration with H2O machine learning frameworks
  • Support for popular data science libraries and tools
  • Collaborative development capabilities
  • Real-time code execution and visualization
  • Support for multiple programming languages, such as Python and R
  • Integration with H2O AI Cloud services
  • Configurable resource allocation and timeout settings
  • Browse, open, and save notebooks in H2O Drive
  • Write and run notebook code with the Enterprise h2oGPTe assistant

Prerequisites​

Before creating a Notebook Engine, make sure you have:

  • Access to H2O AI Cloud web UI. See Access AI Engines for details.
  • Appropriate permissions allocated to your workspace. You need at least the Notebook Engine User role (enginemanager-notebook-engine-user) assigned in the target workspace. See RBAC roles and permissions for the full role breakdown and Engine profile security configuration for how administrators configure profiles and images.
  • (Optional) Resource requirements: Consider your computational needs when configuring resources for your data science workloads. You can also resize a Notebook Engine after you create it.

Create a new Notebook Engine​

To set up a Notebook Engine using the H2O AI Cloud user interface, follow these steps:

Step 1: Navigate to Create AI Engine​

On the AI Engines page, click Create AI Engine.

create ai engine

Step 2: Select Engine Type​

On the Engine Selection tab, select Notebook.

select notebook engine

Step 3: Configure Engine Details​

On the Create AI Engine > Select Engine > Engine Configuration page, provide the following information:

  1. Display Name: Enter a descriptive name for your Notebook engine.

  2. Engine ID: The system automatically generates a unique ID based on your display name. Use an ID that suits your use case. Click the Edit (Pencil icon) icon to customize it.

  3. Profile: Select a Profile from the dropdown. The profile determines the allowed ranges for CPU, GPU, memory, storage, and timeout settings. For more information, see Notebook Engine profiles.

  4. Image: Select the Image from the dropdown. The available images depend on what the administrators have set up. The image also determines which integrations your engine includes — see the H2O Drive integration prerequisites and the Enterprise h2oGPTe assistant prerequisites.

Display Name and Engine ID Guidelines

Display Name:

  • Can include special characters and alphanumeric characters
  • Character limit: 63 characters
  • Use a clear, descriptive name for easy identification

Engine ID:

  • Must start with a lowercase letter
  • Must end with a lowercase letter or digit
  • Can contain only lowercase letters, digits, or dashes (-)
  • Min length: 1 char, Max length: 63 chars
  • Cannot start or end with a dash (-)

Engine Configuration - Display Name and ID settings

Step 4: Configure Resources​

Configure the computational resources for your Notebook engine:

  1. CPU Units: Set the number of CPU units allocated to the engine
  2. GPU Units: Configure GPU units if needed for GPU-accelerated workloads
  3. Memory: Allocate memory in GiB for the notebook environment
  4. Storage: Configure storage capacity in GiB for notebooks and data
Resource ranges are set by your administrator
  • The default values and range for CPU, GPU, Memory, and Storage are set by administrators. Contact your administrator or the H2O support team for more details.
  • Notebook engines support both CPU and GPU workloads for data science and machine learning tasks.
  • Consider your notebook complexity and data size when allocating resources.
  • Resource allocation affects both performance and cost.

Step 5: Configure Shared Access​

Set the Shared toggle to control access to your notebook engine:

  • Turned on: Any user who can access engines in your workspace can open your notebooks, change and delete files in the engine, and run code in it.
  • Turned off: Only you can access the notebook engine.
Sharing Best Practices
  • Shared notebooks allow real-time collaboration and knowledge sharing.
  • You can change sharing settings later by editing the engine when paused.
Shared engines can't use the H2O Drive integration or the Enterprise h2oGPTe assistant

An engine that was created or last resumed with Shared turned on runs without your H2O AI Cloud credentials, so neither the H2O Drive integration nor the Enterprise h2oGPTe assistant can reach your storage or h2oGPTe. Leave Shared turned off if you want to use either one. AI Engine Manager issues the credentials when the engine is created and each time it's resumed, so turning Shared off on a paused engine takes effect when you resume the engine.

Real-time collaboration and these two integrations are mutually exclusive: AI Engine Manager starts JupyterLab with real-time collaboration turned on only when Shared is on, and issues your H2O AI Cloud credentials only when it's off.

Local workspace files survive pausing and resuming, so if you used the H2O Drive integration while the engine wasn't shared, delete local copies of sensitive files before you turn Shared on.

Step 6: Configure Timeout Settings​

Expand the Timeout Configuration section to set:

  1. Max Idle Time (Hours): Set how long the engine can be idle before automatically pausing.
  2. Max Up Time (Hours): Set the maximum duration an engine can run before automatically pausing.
Timeout Best Practices
  • For notebook engines, consider the time needed for data analysis and model development.
  • You can change timeout settings later by editing the engine when paused.

Step 7: Create the Engine​

  1. Review all your configuration settings.
  2. Click Create to create the Notebook engine.
  3. The engine is provisioned and appears in your AI Engines list.

Create Notebook Engine

Feedback