Skip to main content

Configure Kubernetes Cluster

Deployments in the GlobalAI platform run on Kubernetes clusters. Before you can deploy application templates or manage infrastructure resources, you must add your cluster and configure how GlobalAI connects to it.

This guide walks you through adding an existing Kubernetes cluster, configuring secure access, and enabling monitoring for metrics and cost tracking. For an overview of how clusters are used within GlobalAI’s infrastructure model, see Kubernetes Management.

  1. Log in to the GlobalAI platform.
  2. In the left navigation, select Environments.
  3. Go to Cluster Management and select Clusters.

You’ll see a list of predefined clusters. To perform cluster management actions, you must have admin access or explicit permissions for the cluster.

Add a new cluster​

  1. Click Add existing cluster.
  2. Enter a Name for the cluster.
  3. Set the Status to Active.

Your license determines the maximum number of active clusters you can maintain. If you reach this limit, disable an unused cluster before adding a new one.

Configure connection details​

GlobalAI requires credentials to communicate with your Kubernetes cluster. You can connect using a kubeconfig file or cloud-provider credentials.

Use kubeconfig​

This is the standard connection method for most Kubernetes clusters.

  1. In the Configuration section, select Kubeconfig.
  2. Upload your kubeconfig file by dragging and dropping it into the upload area, or paste the raw configuration values directly into the fields.

Use cloud credentials​

For managed Kubernetes services (such as AKS), you can authenticate using cloud-provider credentials.

  1. In the Configuration section, select IAM.
  2. Enter the required credentials to authorize access to the cluster.

Once saved, all credentials are securely stored in Vault and are not displayed in plain text.

Set up monitoring and cost tracking​

You can connect your cluster to external services to monitor performance and track infrastructure costs.

  • Metrics: Enter the Prometheus URL to allow GlobalAI to retrieve performance metrics from the cluster.
  • Cost tracking: Enter the Usage Service URL (Nexus Usage) to enable cost-related metrics. The Nexus Usage service must be installed on the cluster for cost tracking to function.

These integrations support visibility across deployments and cluster usage. For how monitoring data is used at the infrastructure level, see Kubernetes Management.

Validate the connection​

Before saving the configuration:

  1. Click Test connection.
  2. If the test succeeds, click Save to finalize the cluster setup.

Troubleshooting​

Cannot see the cluster list​

If the Cluster Management menu or cluster list is not visible, verify that your user account has the required permissions. You must be an admin or have cluster-level access granted.

License limit reached​

If you are unable to add a new cluster, you may have reached the maximum number of active clusters allowed by your license. Disable an existing cluster to free capacity without deleting it.

Now that you know how to configure a Kubernetes cluster, explore these next steps:

Manage Cluster Groups

Organize clusters and define deployment routing rules.

Deploy Templates

Deploy applications to your configured clusters using templates.

Infrastructure Monitoring

View metrics and cost analysis for your active clusters.