To reduce GPU usage charges over weekends, an administrator decides to temporarily pause all active...

Comprehensive and Detailed Explanation From Exact Extract of Nutanix Artificial Intelligence documents knowledge without any external URL or links:

To temporarily pause an active endpoint and release the associated underlying resources (specifically the GPUs) without deleting the configuration, the administrator should use the Hibernate action.

The Nutanix Enterprise AI dashboard provides specific "Hibernate" and "Resume" actions for endpoints. When an endpoint is hibernated, its status changes toHibernated, and the system scales down the Kubernetes resources, effectively freeing up the GPU for other uses or cost savings. This is the direct administrative action supported in the UI for this goal, whereas "scaling pods to zero" is the underlying technical mechanism (often handled by the serverless framework) rather than the explicit user action button.