Jump to content

Monitoring GFX metrics via CLI or timeseries database?


2 posts in this topic

Recommended Posts

I use Telegraf to collect system metrics in an InfluxDB for monitoring in Grafana. With my Win & Linux systems i'm able to get nvidia GFX metrics via Telegraf's included nvidia_smi component, which does not exist for Mac. FakeSMC/HWMonitor shows me temps, voltages, and fanspeeds of the nvidia cards, and iStat Menus shows those plus processor and memory load for both the nvidia discrete GPU and the intel integrated GPU, but i've yet to find a way to get the metrics they report out to text so I can parse & format them for sending to my timeseries database. Ideally I want dGPU & iGPU processor and memory load, voltages, and dGPU die temperature returned from a command I can run and parse every 5s.

 

Any advice is greatly appreciated.

×
×
  • Create New...