PlantGuard

PlantGuard AI — Production Deployment Guide

PlantGuard AI is structured as an npm monorepo with two primary deployable workspaces:

1. Frontend Client (client/): React 19 + TypeScript + Vite Progressive Web Application (PWA).
2. Backend Server (server/): Node.js 22+ Express REST API with MongoDB and Google Gemini AI.

---

1. System Requirements & Prerequisites

• Node.js: v22.0.0 or higher.
• npm: v10.0.0 or higher (supports npm workspaces).
• Database: MongoDB v6.0+ (Community, Enterprise, or MongoDB Atlas replica set).
• AI Credentials: Google Gemini API Key with access to gemini-3.8-flash.

---

2. Environment Variables

2.1 Server Environment (server/.env)

Copy server/.env.example to server/.env:

# Server Networking
PORT=3000
NODE_ENV=production
CLIENT_ORIGIN=https://plantguard.example.com

# Database Connection (MongoDB)
MONGODB_URI=mongodb+srv://<user>:<password>@cluster.mongodb.net/plantguard_ai?retryWrites=true&w=majority

# Authentication Secrets (generate with `openssl rand -hex 64`)
JWT_SECRET=your_production_access_token_secret_here
JWT_REFRESH_SECRET=your_production_refresh_token_secret_here
JWT_EXPIRES_IN=15m
JWT_REFRESH_EXPIRES_IN=7d

# Google Gemini Vision AI
GEMINI_API_KEY=AIzaSy...
GEMINI_MODEL=gemini-3.8-flash

# Logging
LOG_LEVEL=info

2.2 Client Environment (client/.env)

Copy client/.env.example to client/.env:

# Production API Base URL (leave blank if frontend is served behind reverse proxy /api route)
VITE_API_BASE_URL=https://api.plantguard.example.com

---

3. Build & Pre-Flight Verification

Run the unified monorepo verification pipeline from the repository root:

# 1. Install dependencies across all workspaces
npm install

# 2. Run typecheck, unit tests, and production builds
npm run check

# 3. Ensure MongoDB indexes and migrations are applied
npm --prefix server run db:indexes
npm --prefix server run db:migrate

---

4. Deployment Architectures

Architecture A: Modern Cloud PaaS (Recommended)

• Frontend PWA: Host client/dist/ on Vercel, Netlify, or Cloudflare Pages.
  • Set build command: npm --prefix client run build
  • Output directory: client/dist
  • Configure SPA fallback routing (/* -> /index.html).
• Backend API: Deploy server/ to Render, Railway, Fly.io, or Google Cloud Run.
  • Set build command: npm --prefix server run build
  • Start command: node server/dist/server.js
  • Health check path: /api/v1/health
• Database: MongoDB Atlas (M0 Free tier or dedicated cluster).

---

Architecture B: Docker & Docker Compose

A complete production docker-compose.yml configuration:

version: '3.8'

services:
  mongodb:
    image: mongo:7.0
    restart: always
    volumes:
      - mongo_data:/data/db
    ports:
      - "27017:27017"

  server:
    build:
      context: .
      dockerfile: server/Dockerfile
    restart: always
    environment:
      - PORT=3000
      - NODE_ENV=production
      - MONGODB_URI=mongodb://mongodb:27017/plantguard_ai
      - GEMINI_API_KEY=${GEMINI_API_KEY}
      - GEMINI_MODEL=gemini-3.8-flash
      - JWT_SECRET=${JWT_SECRET}
      - JWT_REFRESH_SECRET=${JWT_REFRESH_SECRET}
      - CLIENT_ORIGIN=https://plantguard.example.com
    depends_on:
      - mongodb
    ports:
      - "3000:3000"

  client:
    build:
      context: .
      dockerfile: client/Dockerfile
    restart: always
    ports:
      - "80:80"
      - "443:443"

volumes:
  mongo_data:

---

5. Health Probes & Monitoring

Configure container orchestrators and load balancers to use these probe endpoints:

• Liveness Probe: GET /api/v1/health
  • Returns HTTP 200 { "status": "ok" } when the Express process is responsive.
• Readiness Probe: GET /api/v1/health/detailed
  • Returns HTTP 200 only if MongoDB connection readyState === 1 and critical subsystems are healthy.

This site is open source. Improve this page.