Monitoring NVIDIA MIG Instances

Hi everyone,

I would like to know the possibilities of NVIDIA MIG instance monitoring.
At the very least I would like to monitor memory usage of an instance, as in nvidia-mig.

Im rather new to using Grafana and Prometheus, so sorry in advance if the question is rather confusing.

Thanks in advance

this might be the route to go?

https://docs.nvidia.com/datacenter/cloud-native/gpu-telemetry/latest/dcgm-exporter.html#:~:text=The%20new%20Multi-Instance%20GPU,seen%20in%20the%20output%20below: