Access a cluster

After your cluster finishes creating, use this page to get its kubeconfig and review the configuration you set at creation time.

1. Open the cluster.

   Go to AI infrastructure → Managed GPU cluster, then select the cluster you want to view.

2. Review Essential properties.

   The tab shows:

   • Cluster name, Version, Network (the subnet selected at creation), and Status.
   • Configuration: select Download cluster configuration to get the kubeconfig you use to connect with kubectl.
   • SSH public key: the key selected at creation, used to SSH into worker nodes.
   • Cluster endpoint access: the access mode (Public/Public & private/Private) and Allow CIDR.
   • SSO / OIDC (beta): shows Disabled unless you've configured it — see Configure OIDC authentication.
   • GPU software information: the GPU software installed on the cluster.
   • Load balancer service: the Internal LB subnet you selected at creation.
   • API: the API URL for the cluster.

3. Review Node pools.

   This tab lists every Worker Group in the cluster and its configuration:

   • Name: the Worker Group name.
   • Is Base: whether this is the non-deletable Base group created with the cluster.
   • Flavor: the GPU server model for this group.
   • Number of servers: how many bare metal GPU servers run in this group.

Next steps

• Configure OIDC authentication — let team members log in with their own SSO account instead of sharing the admin kubeconfig you downloaded above.
• Log in to the cluster with SSO — once OIDC is configured, install kubelogin and assign RBAC for your team.