DEPLOYMENT.md

Deployment Guide

This document outlines the steps required to deploy the Performance Monitor application in a production environment.

Prerequisites

1. Docker: Ensure Docker is installed on the host machine.
2. Docker Compose: Ensure Docker Compose is installed.
3. Configuration: Update the config.yaml file with production server details.

Steps

1. Clone the Repository

Clone the repository to your local machine:

git clone https://github.com/SolidRusT/srt-inference-monitoring.git
cd srt-inference-monitoring

2. Update Configuration

Ensure your config.yaml file is correctly configured with your production server details. If you have different configurations for development and production, create a separate config-prod.yaml and update the Docker Compose file to use it.

3. Build and Run with Docker Compose

Build and run the application using Docker Compose:

docker-compose up -d --build

This command will build the Docker images and start the containers in detached mode.

4. Verify Deployment

Once the containers are up and running, verify the deployment by accessing the dashboard:

http://<your-server-ip>:5000

5. Scaling

To scale the application horizontally, update the docker-compose.yml file to specify the number of replicas:

services:
  app:
    ...
    deploy:
      replicas: 3
    ...

6. Stopping the Application

To stop the application, use the following command:

docker-compose down

This command will stop and remove the containers.

7. Logs and Monitoring

To view the logs of the running application:

docker-compose logs -f

Notes

• Environment Variables: Ensure the necessary environment variables are set for production.
• Data Persistence: Ensure data persistence is properly configured for the Valkey service.

By following these steps, you should be able to deploy the Performance Monitor application in a production environment effectively.