How to Retrieve Cluster Metrics

1 min readUpdated 28 October 2026

Retrieve GPU and fabric performance metrics for a Vultr GPU Fabric Cluster using the Console or API.

Retrieving cluster metrics returns GPU and fabric performance data for all instances in a cluster. This is useful for monitoring cluster health, identifying performance bottlenecks, and tracking resource utilization.

Follow this guide to retrieve cluster metrics using the Vultr Console or the Vultr API.

In the left-hand menu, click Compute, then click Clusters.
Click your target GPU Fabric Cluster to open its management page.
In Metrics, select a tab: Nodes Overview, GPU Memory Utilization, GPU Thermal Throttling, GPU Power Usage, GPU ECC Errors, Fabric Network Utilization, or Fabric Network Errors. Nodes Overview shows each node's CPU Utilization, Memory Util, Memory Swap, and I/O Wait. Click Refresh Data to reload, or toggle Node Name to show or hide node labels.
In Graphs, select an instance from the dropdown under GPU Temperature, Utilization, Power, or Memory, then click an Interval button (5 min, 1 hour, 1 week, 1 month, or 1 year) to reload that graph.

0 Comments

Start the conversation

Be the first to comment and share your perspective with the community.

Retrieve GPU Fabric Cluster Metrics on Vultr Cloud | Vultr Docs