Skip to main content
Version: v1.2.x

Configure System Settings

Use System Settings to manage System-scoped configuration in the PaletteAI UI. System is the top-level scope, backed by the pai-system namespace and managed by platform operators. From this page, you can manage organization-wide configuration such as integrations, GPU limits, Compute Configs, Scaling Policies, Model as a Service Mappings, and Multi-Instance GPU (MIG) availability that cascades to every Tenant and Project.

To modify settings for a single Tenant, refer to Configure Tenant Settings. For more information on the different scopes, refer to System Scope and Multi-Tenancy.

Prerequisites​

  • You must have the System administrator role. For details, refer to Roles and Permissions.

  • System-scoped resources are created in the pai-system namespace. Do not confuse the System scope with the internal platform namespace (mural-system by default), which holds installer-managed resources that cannot be edited in place.

Configure System Settings​

  1. Log in to PaletteAI.

  2. From the Projects Menu in the top-left, select the System scope.

  3. From the left main menu, select System Settings.

Manage System Settings Ref​

A System Settings Ref holds integrations shared with Tenants and Projects in the organization. The System scope has at most one Settings Ref. System-level Settings share Hugging Face and NVIDIA NGC integrations with, and can lock them for, the Tenants and their Projects that you select; by default, an integration is shared with all Tenants and their Projects.

For background on how Settings inheritance works across scopes, refer to Settings and Integrations.

Add an Integration​

Each integration represents a connection to an external system, such as Palette, Hugging Face, or NVIDIA NIMs. Each integration type can be added once per System Settings Ref.

  1. On the Settings Ref tab, on the System sub-tab, select Add Integration to open the Add Integration drawer.

  2. Choose an Integration Type — Palette, Hugging Face, or NVIDIA NIMs. The form fields update to match the selected type.

  3. Complete the fields for the chosen integration type:

    • Palette integration — used to provision compute resources from your Palette tenant. Provide the Integration Name, Host URL, Project ID, and API Key. Select Validate to verify the credentials against the Palette server.
    • Hugging Face integration — enables Hugging Face model management features. Provide the Integration Name and your Hugging Face Hub API Key. Select Validate to verify the key against the Hugging Face API.
    • NVIDIA NIMs integration — enables NVIDIA NGC model management features. Provide the Integration Name and your NVIDIA NGC API Key. Select Validate to check that a key is provided. The Validate action does not verify the key format or contact NVIDIA NGC.

  4. (Optional) Expand Metadata to add Kubernetes labels and annotations to the integration.

  5. (Hugging Face and NVIDIA NIMs only) Under Shared access, choose which Tenants and Projects can use this integration, and optionally lock the integration so lower scopes cannot override it.

  6. Select Confirm to add the integration.

Edit or Delete an Integration​

Use the action menu on each row of the Integrations table to edit or delete an integration. Palette integrations cannot be deleted; use Edit & Share to rotate their credentials.

Manage Resource Limits​

On the Resource Limits tab, set organization-wide GPU limits and requests that act as ceilings on every Tenant, and review or adjust per-Tenant GPU quotas. For the full quota model, refer to GPU Quotas.

  1. On the System sub-tab, update the system-wide GPU Limits and GPU Requests, and then select Save changes.

    Each entry must use a unique GPU variant and a limit value. If a warning indicates that Tenants exceed the system GPU limits, increase the system limits or reduce Tenant limits.

  2. Select the Tenants sub-tab to review GPU limits and requests across Tenants.

  3. Search or filter to locate the Tenant, and then select Edit to adjust that Tenant's GPU limits and requests.

Manage Multi-Instance GPU​

Govern where Multi-Instance GPU (MIG) configuration appears in the user interface across the whole organization. System decisions are authoritative: neither a Tenant nor a Project can override the System scope's MIG governance. These controls affect only where MIG configuration is exposed; they do not change how the NVIDIA GPU Operator applies MIG.

  1. Open the System MIG settings.

  2. Enable or disable MIG configuration across all Tenants and Projects.

  3. Choose which Tenants may configure MIG in their Compute Pool flows.

    • Allow all Tenants to expose MIG configuration. This is the default.
    • Restrict MIG configuration to a specific list of Tenants.
    • Allow no Tenants to expose MIG configuration.

  4. Save your changes.

Tenants further refine which of their Projects may configure MIG. Refer to Manage Multi-Instance GPU in Tenant Settings for the Tenant-scoped controls.

info

MIG configuration also depends on the hardware. Even when the System scope allows MIG, PaletteAI offers MIG configurations only for Compute Pools whose GPUs support MIG. Refer to Multi-Instance GPU for the hardware requirements.

Create Compute Config​

Create a System-scoped Compute Config that Tenants and Projects can adopt as a shared infrastructure baseline. Each Compute Config captures cluster defaults such as Secure Shell (SSH) keys, Network Time Protocol (NTP) servers, and the network overlay.

  1. On the Compute Config tab, on the System sub-tab, select Create Compute Config and complete the wizard.

  2. Under Shared access, choose which Tenants and Projects can use this Compute Config.

  3. Review the configuration and select Create.

Edit, clone, or delete a System Compute Config from the action menu on each row. Deletion is blocked while a Compute Config is in use.

Create Scaling Policy​

Create a System Scaling Policy that Compute Pools across the organization can reference to autoscale workloads.

For background on autoscaling and the YAML reference, refer to Create and Manage Scaling Policies.

  1. On the Autoscaling Policies tab, on the System sub-tab, select Create Autoscaling Policy and provide the policy name, metrics, resource bounds, cooldown, and timeout.

  2. Under Shared access, choose which Tenants and Projects can use this policy.

  3. Select Create to save the policy.

Edit, clone, or delete a System Scaling Policy from the action menu on each row. Deletion is disabled while the policy is referenced by any Compute Pool.

Create Model as a Service Mappings​

Create a mapping on the Model as a Service Mappings tab to route incoming Model Deployments to a chosen Profile Bundle based on the model's source and attributes. When a Model Deployment matches the source and filters defined for a mapping, PaletteAI deploys the model using the target Profile Bundle.

  1. On the System sub-tab, select Add Mapping, choose the model source and match filter, and pick the target Profile Bundle.

  2. Select Confirm to save the mapping.

Edit or delete an existing mapping from the action menu on each row.

Validate​

  1. Refresh the System Settings page and confirm your saved changes persist on the applicable System or Tenants sub-tab.

Next Steps​