# utkuozdemir/nvidia_gpu_exporter

Nvidia GPU exporter for prometheus using nvidia-smi binary OR using NVML

Repository: https://github.com/utkuozdemir/nvidia_gpu_exporter
Canonical: https://ross.abutalabs.com/products/nvidia_gpu_exporter
Language: Go
License: MIT
License Family: permissive
Topics: prometheus, prometheus-exporter, nvidia-gpu, nvidia, nvidia-smi, monitoring, ai, cryptocurrency, gaming, llm, llm-training, inference, nvml
Last push: 2026-08-25T23:58:09+00:00

## Health v2 (maintenance only)
Score: 98/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 97, longevity 100
- inputs: {"age_days": 1912, "days_push": 8, "days_rel": 21, "gap_med": 4.5, "n_releases_24m": 17}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1542, forks 154 (observed 2026-08-28T04:05:01.074960+00:00)

## What it is
A Prometheus exporter written in Go that collects NVIDIA GPU metrics by invoking the nvidia-smi binary or reading directly from the NVML driver library. It works on Windows, Linux, and macOS without C bindings and can even run remotely from the monitored machine.

## Use cases
- monitor nvidia gpu utilization memory power and temperature in prometheus
- export gpu metrics from windows machines without docker or linux
- scrape gpu stats for gaming rigs on a grafana dashboard
- monitor gpus in small kubernetes clusters or homelabs without the gpu operator
- collect gpu metrics in vgpu mig or restricted container environments
- monitor mixed fleets of old and new nvidia cards with one exporter
- scrape nvidia-smi metrics remotely over ssh

## When to choose
- you have consumer or prosumer GeForce/RTX cards where nvidia-smi is the only uniform metric source
- you need GPU metrics on Windows or in restricted environments where DCGM is unavailable
- you want a lightweight exporter without C bindings or the NVIDIA GPU Operator stack
- you need one exporter that behaves consistently across old and new cards

## When to avoid
- you run datacenter GPUs on Kubernetes with the GPU Operator already installed - use DCGM-exporter instead
- you need deep GPU counters or profiling metrics beyond what nvidia-smi/NVML expose
- you need a vendor-supported solution with guaranteed maintenance

## Facets
- artifact type: service
- maturity: active
- function: monitoring
- domain: monitoring, gpu-computing
- platform: windows, self-hosted
- tags: prometheus-exporter, nvidia-gpu, nvidia-smi, nvml, gpu-metrics, homelab, devops, linux, macos, docker

## Member repositories
- utkuozdemir/nvidia_gpu_exporter (main) score 98

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:01.074960+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:30:46.029092+00:00, confidence not recorded.
  - readme: https://github.com/utkuozdemir/nvidia_gpu_exporter (fetched 2026-08-28T04:05:01.074960+00:00, sha dc4cf3194a5c)
- Data as of 2026-08-30T08:39:29.467469+00:00.