Skip to content

Managing Workspaces

A Workspace represents a user-driven environment (a “job”) for interactive data science or development—often running tools like Jupyter Notebook or VS Code. Each Workspace is powered by a Docker image, assigned specific resources (GPU, CPU, Memory), and optionally linked to persistent volumes for storage. Users can also enable checkpointing to periodically capture the state of their session for recovery purposes. This approach ensures that development and experimentation can be conducted in a consistent, easily reproducible environment.

Using Workspaces, your team can seamlessly develop and experiment with AI models, data processing scripts, or any other workload, all within a controlled, resource-managed environment.

Accessing the Workspaces Dashboard

  • In the left navigation bar, select the Workspaces Icon Workspaces Icon .

  • You will land on the Workspaces dashboard, which shows all current Workspaces.

    Workspace Dashboard

Workspaces Dashboard Elements

This section describes the key elements found on the Workspaces dashboard, which provides an overview and management interface for all active and configured workspaces within MemVerge.AI.

Search and Filter Controls

  • Search Workspaces Field: An input field with a magnifying glass icon, allowing users to search for specific workspaces by name.

  • All Projects Dropdown: A filter to display workspaces belonging to a specific project or all projects.

  • Workspace Status Summary: A concise overview of workspace states:

    • All: Total count of all workspaces.
    • Ready: Number of workspaces that are ready for use.
    • NotReady: Number of workspaces that are not yet ready or have issues.
    • Stopped: Number of workspaces that are currently stopped.

Display Options and Actions

  • View Toggle Icons: Icons to switch between different display layouts (e.g., grid view and list view).
  • + Create Workspace Button: A button to initiate the process of creating a new workspace.

Workspace Card (Example: red-ws)

Each card represents an individual workspace and displays its key details and quick actions:

  • Workspace Icon: A visual icon representing the workspace type (e.g., Jupyter Notebook icon).

  • Workspace Name: The name of the workspace (e.g., red-ws).

  • Status Tag: A visual tag indicating the current status of the workspace (e.g., Ready, Running, Stopped).

  • Image/Source: Displays the container image used for the workspace (e.g., Image: ghcr.io/memverge/mvaiwork...).

  • Resource Summary: Key resources allocated to the workspace:

    • GPU: Number of GPUs.
    • CPU Cores: Number of CPU cores.
    • MEM: Amount of memory (in GiB).

  • Current Activity/State: Information about ongoing processes within the workspace (e.g., Checkpointing).

  • Age: The duration since the workspace was started or last updated (e.g., Age: 5h 34m 17s).

  • Action Buttons: Quick actions directly accessible from the card:

    • Stop: To halt the workspace.
    • Connect: To establish a connection to the workspace.
    • Delete: To permanently remove the workspace.

Pagination Controls (Bottom Right)

  • Navigation Arrows: To move between pages of workspace listings.
  • Items Per Page Selector: Indicates the current page and allows adjustment of the number of workspaces displayed per page (e.g., 1 / 20 / page).

Creating, Modifying, and Deleting Workspaces

To manage Workspaces on bahalf of users, you can refer to the following sub-pages: