AI is taking over in our datacenter now. Many developers and users are interested in monitoring the regular things we monitor with server ie. CPU/Memory/Storage etc. but now GPU's from Nvidia are quickly becoming part of the infrastructure. These can be monitored today through API https://github.com/NVIDIA/go-nvml or using https://docs.nvidia.com/deploy/nvidia-smi/index.html but it would be nice if Observability could do this natively inside Observability On Prem Edition. If you have an AI cluster of server all loaded with Nvidia GPU's AI developers and users would like to be able to monitor their GPU usage along with all of the other node details.
