Nvidia GPU Monitor

Posted Nov 24, 2024 Updated Nov 25, 2024

By bigsk1

6 min read

GPU Monitor: Your NVIDIA GPU’s Best Friend 🎮

Ever wondered what your GPU is up to? Whether you’re training ML models, rendering videos, or gaming hard, keeping an eye on your GPU’s health is crucial. Enter GPU Monitor - a sleek, real-time dashboard that turns boring metrics into beautiful visualizations.

Who Needs This?

ML Engineers training models for days (Is that GPU throttling?)
Content Creators rendering their masterpieces (Why is this export taking so long?)
Gamers pushing their hardware to the limit (Is it time to dial back those settings?)

Why You’ll Love It

Real-time Metrics: Temperature, utilization, memory usage, and power consumption at a glance
Historical Data: See how your GPU performs over time (15min to 24hr views)
Smart Alerts: Get notified when your GPU needs attention
Clean Interface: Modern, dark theme dashboard that’s easy on the eyes
Lightweight: Uses minimal resources while monitoring
Docker-based: Run it anywhere, no messy installations

Quick Setup

Got Docker? You’re halfway there! Just run:

  
docker run -d \
  --name gpu-monitor \
  -p 8081:8081 \
  -e TZ=America/Los_Angeles \
  -v /etc/localtime:/etc/localtime:ro \
  -v ./history:/app/history \
  -v ./logs:/app/logs \
  --gpus all \
  --restart unless-stopped \
  bigsk1/gpu-monitor:latest

Open http://localhost:8081 in your browser
Watch your GPU metrics come alive!

Cool Features You Might Not Expect

Clickable Gauges: Toggle different metrics on the performance graph
Intelligent Alerts: Set custom thresholds for temperature, utilization, and power
Sound Notifications: Hear when your GPU needs attention (customizable)
Browser Notifications: Stay informed even when the tab isn’t active
Data Persistence: Track performance across restarts
Mobile Responsive: Check your GPU from your phone

Real-world Uses

For ML Engineers

Track GPU utilization during training runs. Is your model actually using the GPU efficiently? Are you thermal throttling? GPU Monitor helps you optimize your training pipelines.

For Content Creators

Monitor rendering jobs without constantly checking task manager. Get notified when renders complete or if your GPU is struggling.

For Gamers

Keep an eye on temperatures during intense gaming sessions. Perfect for overclocking experiments or optimizing game settings.

Pro Tips

Set custom alert thresholds based on your GPU’s specs
Use browser notifications for headless setups
Check historical data to identify performance patterns
Monitor power efficiency for optimal workloads

Join the Community

Found a bug? Have a feature request? Want to contribute? Check out our GitHub repository!

Remember: A watched GPU never throttles! 🚀

Github Readme

Nvidia GPU Dashboard

A real-time lightweight NVIDIA GPU monitoring dashboard built with Docker for easy deployment and cross-platform compatibility.

Features

Real-time GPU metrics monitoring
Interactive web dashboard
Historical data tracking (15m, 30m, 1h, 6h, 12h, 24h)
Temperature, utilization, memory, and power monitoring
Docker-based for easy deployment
Persist history between new containers
Real time alerts - sound and notification
Responsive theme for any size screen
Toggle gauges on or off to show metrics in graph

Prerequisites

Docker
NVIDIA GPU
NVIDIA Container Toolkit

Quick Start

Test to see if you already have the requirements and ready to use.

  
sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi

If this failed proceed to Installation Prerequisites

Using Pre-built Image

  
docker run -d \
  --name gpu-monitor \
  -p 8081:8081 \
  -e TZ=America/Los_Angeles \
  -v /etc/localtime:/etc/localtime:ro \
  -v ./history:/app/history \
  -v ./logs:/app/logs \
  --gpus all \
  --restart unless-stopped \
  bigsk1/gpu-monitor:latest

Note: Update your timezone to use the correct time

Using Docker Compose

Clone the repository:

git clone https://github.com/bigsk1/gpu-monitor.git
cd gpu-monitor

Start the container:
1 docker-compose up -d
Access the dashboard at: http://localhost:8081

Installation Prerequisites

1. Ubuntu / Debian / WSL

Windows users make sure you have wsl with docker an easy way is Docker Desktop Installation for Windows

Installing with apt add NVIDIA package repositories

  
curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list

2. Install nvidia container toolkit

sudo apt-get update

  
sudo apt-get install -y nvidia-container-toolkit

3. Configure Docker with toolkit

  
sudo nvidia-ctk runtime configure --runtime=docker

The nvidia-ctk command modifies the /etc/docker/daemon.json file on the host. The file is updated so that Docker can use the NVIDIA Container Runtime.

4. Restart Docker daemon

sudo systemctl restart docker

5. Test to see if installed correctly

  
sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi

For other distributions, check the official documentation.

Building gpu-monitor from source

Clone the repository:

git clone https://github.com/bigsk1/gpu-monitor.git
cd gpu-monitor

Build the image:
1 docker build -t gpu-monitor .

Run the container:

  
docker run -d \
  --name gpu-monitor \
  -p 8081:8081 \
  -e TZ=America/Los_Angeles \
  -v /etc/localtime:/etc/localtime:ro \
  -v ./history:/app/history \
  -v ./logs:/app/logs \
  --gpus all \
  --restart unless-stopped \
  gpu-monitor

Configuration

The dashboard is accessible at: http://localhost:8081 by default. To change the port, modify the docker-compose.yml file or the -p parameter in the docker run command.

Alternative Setup Method

A setup script is provided for convenience. It checks prerequisites and manages the service:

If you have issues then make sure setup.sh is executable

chmod +x ./setup.sh

Check prerequisites and start the service
1 ./setup.sh start
Stop the service
1 ./setup.sh stop
Restart the service
1 ./setup.sh restart
Check service status
1 ./setup.sh status
View logs
1 ./setup.sh logs

Example of script running

  
~/gpu-monitor ./setup.sh start
[+] Checking prerequisites...
[+] Docker: Found
[+] Docker Compose: Found
[+] NVIDIA Docker Runtime: Found
[+] NVIDIA GPU: Found
[+] Starting GPU Monitor...
Creating network "gpu-monitor_default" with the default driver
Creating gpu-monitor ... done
[+] GPU Monitor started successfully!
[+] Dashboard available at: http://localhost:8081
[+] To check logs: docker-compose logs -f

Data Persistence

By default, all data is stored within the container will persist between container rebuilds, if you don’t want that then remove volumes, modify the docker run or docker-compose.yml:

  
services:
  gpu-monitor:
    # ... other settings ...
    volumes:
      - ./history:/app/history    # Remove Persist historical data
      - ./logs:/app/logs    # Remove Persist logs

Alerts

You can enable or disable alerts in ui, you can set thresholds for gpu temp, gpu utilization % and watts. Setting are saved in your browser if you make changes you only need to do it once, however you can always modify the code and rebuild the container to make it permanent.

The defaults are:

    temperature: 80,  
    utilization: 100,
    power: 300

Troubleshooting

Common Issues

NVIDIA SMI not found
- Ensure NVIDIA drivers are installed
- Verify NVIDIA Container Toolkit installation
- Make sure you can run:

  
sudo docker run --rm --runtime=nvidia --gpus all ubuntu nvidia-smi