nvtop-exporter 0.0.7

Add your description here

nvtop exporter

Prometheus exporter for syllo/nvtop

Metrics

# HELP gpu_clock_mhz gpu clock MHz
# TYPE gpu_clock_mhz gauge
gpu_clock_mhz{device="DG2 (Arc A770)",index="0"} 2400.0
# HELP gpu_mem_clock_mhz mem clock MHz
# TYPE gpu_mem_clock_mhz gauge
gpu_mem_clock_mhz{device="DG2 (Arc A770)",index="0"} 0.0
# HELP gpu_temp_celsius gpu temp celsius
# TYPE gpu_temp_celsius gauge
gpu_temp_celsius{device="DG2 (Arc A770)",index="0"} 48.0
# HELP gpu_fan_speed_rpm gpu fan speed rpm
# TYPE gpu_fan_speed_rpm gauge
gpu_fan_speed_rpm{device="DG2 (Arc A770)",index="0"} 302.0
# HELP gpu_power_draw_watts gpu power draw watts
# TYPE gpu_power_draw_watts gauge
gpu_power_draw_watts{device="DG2 (Arc A770)",index="0"} 38.0
# HELP gpu_usage gpu utilization %
# TYPE gpu_usage gauge
gpu_usage{device="DG2 (Arc A770)",index="0"} 0.0
# HELP gpu_mem_usage gpu memory utilization %
# TYPE gpu_mem_usage gauge
gpu_mem_usage{device="DG2 (Arc A770)",index="0"} 0.0
# HELP gpu_encode_decode_usage gpu encode/decode utilization %
# TYPE gpu_encode_decode_usage gauge
gpu_encode_decode_usage{device="DG2 (Arc A770)",index="0"} 0.0
# HELP gpu_mem_total_bytes gpu memory total bytes
# TYPE gpu_mem_total_bytes gauge
gpu_mem_total_bytes{device="DG2 (Arc A770)",index="0"} 1.7079205888e+010
# HELP gpu_mem_used_bytes gpu memory used bytes
# TYPE gpu_mem_used_bytes gauge
gpu_mem_used_bytes{device="DG2 (Arc A770)",index="0"} 7.3265152e+07
# HELP gpu_mem_free_bytes gpu memory free bytes
# TYPE gpu_mem_free_bytes gauge
gpu_mem_free_bytes{device="DG2 (Arc A770)",index="0"} 1.7005940736e+010
# HELP gpu_process_usage gpu process utilization %
# TYPE gpu_process_usage gauge
gpu_process_usage{cmdline="/usr/lib/plexmediaserver/Plex Transcoder ...",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1522258",user="trev"} 0.0
gpu_process_usage{cmdline="nvtop -s",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1550252",user="root"} 0.0
# HELP gpu_process_mem_usage gpu process memory utilization %
# TYPE gpu_process_mem_usage gauge
gpu_process_mem_usage{cmdline="/usr/lib/plexmediaserver/Plex Transcoder ...",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1522258",user="trev"} 0.01
gpu_process_mem_usage{cmdline="nvtop -s",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1550252",user="root"} 0.0
# HELP gpu_process_encode_decode_usage gpu process encode/decode utilization %
# TYPE gpu_process_encode_decode_usage gauge
gpu_process_encode_decode_usage{cmdline="/usr/lib/plexmediaserver/Plex Transcoder ...",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1522258",user="trev"} 0.0
gpu_process_encode_decode_usage{cmdline="nvtop -s",device="DG2 (Arc A770)",index="0",kind="graphic",pid="1550252",user="root"} 0.0

Use

$ ./nvtop-exporter
2026-02-23 11:52:09,536 - INFO - Starting server on :8080

Environment

Variable	Default	Description
PORT	8080	Port for the HTTP server to listen on
LOG_LEVEL	INFO	How verbose the logs should be

Prometheus

scrape_configs:
  - job_name: nvtop
    static_configs:
      - targets:
          - 127.0.0.1:8080"

Install

AppImage

nvtop-exporter_0.0.7_linux_amd64.AppImage

$ chmod +x nvtop-exporter_0.0.7_linux_amd64.AppImage
$ ./nvtop-exporter_0.0.7_linux_amd64.AppImage
2026-02-23 11:52:09,536 - INFO - Starting server on :8080

Service

nvtop_exporter.service

[Unit]
Description=nvtop gpu exporter
After=network.target

[Service]
ExecStart=/usr/local/bin/nvtop-exporter_0.0.7_linux_amd64.AppImage
Type=simple
Restart=on-failure

[Install]
WantedBy=multi-user.target

Docker

docker run ghcr.io/spotdemo4/nvtop-exporter:0.0.7

Nvidia

docker-compose.yaml

services:
  nvtop-exporter:
    image: ghcr.io/spotdemo4/nvtop-exporter:0.0.7
    pid: host
    ports:
      - "8080:8080"

    # Expose the GPU
    runtime: nvidia
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities:
                - gpu

Intel

docker-compose.yaml

services:
  nvtop-exporter:
    image: ghcr.io/spotdemo4/nvtop-exporter:0.0.7
    pid: host
    ports:
      - "8080:8080"

    # Expose the GPU
    devices:
      - /dev/dri:/dev/dri
    cap_add:
      - CAP_PERFMON

nix

nix run github:spotdemo4/nvtop-exporter

uv

uvx git+https://github.com/spotdemo4/nvtop-exporter