As giveen1 said in theory it should work on Windows if you have Python3 and NVML (the Nvidia Management Library) installed. I have not tested any of this personally, so if you're up for it I would love to hear feedback whether it works or not.

NVML is part of the CUDA toolkit and can be downloaded from here: https://developer.nvidia.com/cuda-downloads

Thanks

Re multiple servers - As Prometheus is pull based what this script does is simply expose telemetry data as an http endpoint (in Prometheus' data model semantic, ie. a simple specifically formatted text file) so that Prometheus can query it. This means that the script would have to run on the rigs themselves and then you'd configure Prometheus to pull from all rigs. This is safe as all data has the GPU's uuid as a label which means that each GPU will have its own time series stored independent of one another.

Hey all,

I created a small script to extract Nvidia and DSTM telemetry data into Prometheus to create nice dashboards.

Check it out and let me know what you think: https://github.com/pwm/rig-stats