Open monitoring for an instance
From your instance’s details page, click Monitoring. The view is scoped to that instance and lists each GPU attached to it.What you can see
Metrics refresh automatically about every 30 seconds.
Instances without monitoring
Monitoring depends on the platform agent being present on the instance. If an instance’s details page has no Monitoring button, that machine type doesn’t support it yet;nvidia-smi on the instance gives you the same numbers on demand — see Managing Instances.
Telemetry Hyperbolic collects
On virtual machine instances, Hyperbolic collects a small set of health metrics from inside the VM. We use them to spot failing GPUs and other hardware faults sooner, and to troubleshoot faster when you contact support. Collection is on by default for virtual machine instances created on or after October 19, 2026, and you can turn it off at any time. Instances created before then don’t collect.
Metrics leave the VM over mutual TLS with a certificate issued to that instance alone, and are kept for 30 days. Metrics carry the names of the disks, the root filesystem’s device (such as
/dev/vda1), and the network interfaces they describe. The totals are for the whole VM, so they include the resources your workloads use. Collection starts when the VM boots. It doesn’t apply to bare-metal instances or Managed Kubernetes nodes, and some instance types don’t support it yet.
Opting out doesn’t change your instance, its performance, or its price. If you later report a GPU problem, support may ask you for more detail, such as the nvidia-smi -q output.
Opting out
You have root on your VM, so you can opt out from inside it. Create the opt-out file and stop the two collectors:/etc/forge/monitor/disabled exists, neither collector starts again, including after a reboot. Hyperbolic never removes or changes an opt-out file that holds your own text, such as the line in the example above.
To opt back in, remove the file and start the collectors:
What monitoring is not
Monitoring is a read-only view. It does not alert you, and it does not restart or replace a GPU automatically. If you see a GPU that has dropped out or is throttling, run the checks in Verifying Instance Performance and contact support with the instance ID and thenvidia-smi -q output.