Users Overview#
The AMD enterprise AI reference stack has a robust user management system bundled with the authentication system Keycloak. The platform can also use an identity provider for user authentication.
Tasks that require the Keycloak admin console share a common Keycloak admin console access guide. Day-to-day user management in the AMD Resource Manager UI is covered in Manage users.
To manage users on the platform, navigate to the Users page.
Roles#
AMD Resource Manager and AMD AI Workbench have two different user roles: platform admin and team member.
The platform admin has access to everything on the platform. When submitting workloads, they follow the regular project rules, meaning they can submit workloads only to clusters where they have a project. Administrators can change a user’s role.
The team member has several limitations in their workflows. A team member can:
view clusters if any of their projects are in the cluster
view the kubeconfig for the clusters they have access to
view the settings including quotas, storages, and secrets for the projects they have access to
submit workloads to their projects
view workloads in their projects
delete workloads from their projects.
Adjusting user roles#
Roles are currently fixed and cannot be adjusted in individual platform installations.