Skip to content

Deployment.md

Latest commit

 

History

History
73 lines (57 loc) · 2.79 KB

Deployment.md

File metadata and controls

73 lines (57 loc) · 2.79 KB
Raw

Production Deployment Guide

This guide details the steps to deploy the UIDAI Anomaly & Risk Intelligence System in a production environment using best practices. The backend API server and frontend dashboard are split to enable independent scaling and isolation.

🏗️ 1. Backend REST API Deployment (Render / Cloud Docker)

The backend is packaged inside a multi-stage Docker container running FastAPI with Uvicorn.

Deployment Steps on Render:

  1. Push the repository to GitHub.
  2. Sign in to Render.
  3. Click New + and select Web Service.
  4. Connect your GitHub repository.
  5. In the settings, configure:
    • Runtime: Docker
    • Dockerfile Path: Dockerfile (automatically uses backend-only multi-stage build)
    • Instance Type: Free/Paid (as appropriate)
  6. Render will automatically build the Docker image and pass a dynamic PORT environment variable to the container.
  7. Set up the Health Check Path in the Advanced Settings:
    • Path: /health or /api/health
    • Interval: 10-30 seconds
  8. Verify Deployed URL: The service will expose an endpoint like: https://uidai-risk-api.onrender.com

🎨 2. Dashboard Web UI Deployment (Streamlit Community Cloud)

The interactive dashboard runs on Streamlit Community Cloud and communicates with the backend API over HTTP.

Deployment Steps on Streamlit Cloud:

  1. Log in to Streamlit Share.
  2. Click New App.
  3. Choose your GitHub repository, branch (main), and set the main file path to: app.py
  4. Expand the Advanced Settings before deploying.
  5. Under Secrets, define the environment variable to point to your backend API hosted on Render: 
    API_URL = "https://uidai-risk-api.onrender.com"
  6. Click Deploy.
  7. Verify Deployed URL: The service will expose an app like: https://uidai-risk-dashboard.streamlit.app

🎛️ 3. Environment Variables Summary

Variable Name Required By Description Example Value
PORT FastAPI (Render) Port for the backend API container (Render injects this dynamically). 8000
API_URL Streamlit Cloud URL of the running FastAPI server. Streamlit uses this to fetch data. If empty, falls back to local CSV. https://uidai-risk-api.onrender.com

🐋 4. Local Orchestration (Docker Compose)

For local development, you can spin up the backend API using Docker Compose:

docker-compose up --build

This builds and launches the API container on http://localhost:8000. You can run the Streamlit dashboard locally pointing to it:

$env:API_URL="http://localhost:8000"
streamlit run app.py

Or simply run both locally using the orchestrator script:

python start_services.py