Skip to main content
Version: 25.3

Custom container template

Prerequisites

You will need the following to get started:

  • Valid credentials for accessing cloud storage resources
  • Maintain role permissions (minimum)
  • A compute environment with sufficient resources (scale based on data volume)
  • Data Explorer enabled

Select Custom container template and provide your own template (see Custom container template image). When you select this option, you cannot Install Conda packages.

Configure the following fields:

  • Container identifier: The template for the container.
  • Resource labels: Any resource label already defined for the compute environment is added by default. Additional custom resource labels can be added or removed as needed.
  • Environment variable: Environment variables for the session. All variables from the selected compute environment are automatically inherited and displayed. Additional session-specific variables can be added. Session-level variables take precedence. To override an inherited variable, define the same key with a different value.
  • Studio name: The name for the Studio.
  • Description (optional): A description for the Studio.
  • Collaboration: Session access permissions. By default, all workspace users with the launch role and above can connect to the session. Toggle Private on to restrict connections to the session creator only.
    note

    When private, workspace administrators can still start, stop, and delete sessions, but cannot connect to them.

  • Session lifespan: The duration the session remains active. Available options depend on your workspace settings:
    • Stop the session automatically after a predefined period of time: An automatic timeout for the session (minimum: 1 hour; maximum: 120 hours; default: 8 hours). If a workspace-level session lifespan is configured, this field cannot be edited. Changes apply only to the current session and revert to default values after the session stops.
    • Keep the session running: Continuous session operation until manually stopped or an error terminates it. The session continues consuming compute resources until stopped.

Mount data​

Mount data repositories to make them accessible in your session:

  1. Select Mount data to open the data selection modal.
  2. Choose the data repositories to mount.
  3. Select Mount data to confirm.

Mounted repositories are accessible at /workspace/data/<DATA_REPOSITORY> using the Fusion file system. Data doesn't need to match the compute environment region, though cross-region access may increase costs or cause errors.

Sessions have read-only access to mounted data by default. Enable write permissions by adding AWS S3 buckets as Allowed S3 Buckets in your compute environment configuration.

Files uploaded to a mounted bucket during an active session may not be immediately available within that session.

Save and start​

  1. Review the configuration to ensure all settings are correct.
  2. Save your configuration:
    • To save and immediately start your Studio, select Add and start.
    • To save but not immediately start your Studio, select Add only.

Studios you create will be listed on the Studios landing page with a status of either stopped or starting. Select a Studio to inspect its configuration details.