Skip to main content

Monitoring Notification Server

This is the easiest way to deploy your Groundlight detectors on a linux computer. All configuration is done through a web user interface, and no code development is required.

Prerequisites​

  1. Internet- connected linux computer
  2. Video source (USB camera or RTSP stream)
  3. Groundlight API Key (available from groundlight.ai)

Using the Application​

Our Monitoring Notification Server is a server you can deploy anywhere to easily build Groundlight Detectors, and configure them to pull from custom image sources and post notifications.

The Monitoring Notification Server has a simple web interface (depected below) that allows you to configure your detector(s), and a backend that runs on your device to pull images from your camera and post notifications.

MNS Home Page​

Intro Page

Detector Dashboard​

Detector Dashboard

Running the server​

There are several ways to deploy the code:

  • Using Docker Compose
  • Using AWS Greengrass
  • Using Kubernetes

Running with Docker Compose​

  1. Use the file docker-compose.yml.

  2. Run docker-compose up in the same directory as the docker-compose.yml file.

If you're using Docker Compose v2, replace docker-compose with docker compose.

Running from Docker Compose on 32-bit ARM (armv7)​

32-bit arm requires different binary images.

  1. Use the slightly different docker-compose-armv7.yml.

  2. Run docker-compose -f docker-compose-armv7.yml up.

If you're using Docker Compose v2, replace docker-compose with docker compose.

Running with AWS Greengrass​

Before creating the component, run sudo usermod -aG docker ggc_user on your Greengrass device to allow the Greengrass service to access the host's Docker daemon.

  1. Create a new Greengrass Component
  2. Select "Enter recipe as YAML"
  3. Paste the YAML from greengrass-recipe.yaml into the text box
  4. Click "Create component"
  5. Click "Deploy" to deploy the component to your Greengrass group

Running with Kubernetes​

We recommend a minimal Kubernetes install like k3s.

Use kubernetes.yaml manifest.

  1. Create a Kubernetes cluster and install kubectl on your machine.
  2. Run kubectl apply -f kubernetes.yaml in the same directory as the kubernetes.yaml file.

Building from Source​

  1. Install Node.js and Python 3.8+.
git clone https://github.com/groundlight/monitoring-notification-server
cd monitoring-notification-server
npm install
npm run dev

Open http://localhost:3000 with your browser to see the result.

The FastApi server will be running on http://0.0.0.0:8000 – feel free to change the port in package.json (you'll also need to update it in next.config.js).