Resource usage using Grafana

Monitor the resource usage in your IBM Watson® Machine Learning Accelerator cluster using the visual charts available in Grafana.

Before using Grafana

Before using Grafana, consider the following:
  • The Grafana dashboard for additional scheduling metrics can only be viewed by a WML Accelerator cluster administrator. The WML Accelerator cluster administrator has access to Grafana as a Grafana administrator. Only the Grafana administrator role is supported.
  • Before using Grafana for the first time, as a Watson Machine Learning Accelerator cluster administrator, you must run the following command to label your namespace with namespace=openshift-monitoring:
    oc label namespace wmla-namespace namespace=openshift-monitoring
    where wmla-namespace is the name of the Watson Machine Learning Accelerator namespace. This will configure monitoring services required by Grafana.

Log in to Grafana

The Grafana console can be accessed from the Watson Machine Learning Accelerator service console.

To locate and log in to Grafana:
  1. Log in to the Watson Machine Learning Accelerator console as a project administrator. To locate the IBM Watson Machine Learning Accelerator console, see Access the console.
  2. In the upper right, click Username > Grafana to start Grafana.
  3. Log in to Grafana using the default admin account.
    To obtain the password, run:
    oc get cm wmla-grafana-ini -n <wmla-namespace> -o yaml | grep admin_password
    For example:
    # oc get cm wmla-grafana-ini -n wmla -o yaml | grep admin_password
        admin_password = password
    After logging in for the first time, consider changing your password.

Changing your Grafana password

After logging into Grafana, change your password. For details see, Manage user preferences.

Create a visual dashboard in Grafana

Create a dashboard in Grafana that uses your resource usage information:
  1. From the navigation menu, click Create > Dashboard.
  2. Select Add Query.
  3. When creating your query, input one of the following metrics: spectrum_scheduler_consumer_requested_resource_counter or spectrum_scheduler_consumer_used_resource_counter and press Enter on your keyboard.
  4. In the upper right menu bar, click Save Dashboard to save your new dashboard.