Skip to content

DEPLOY.md

Latest commit

 

History

History
230 lines (171 loc) · 6.72 KB

DEPLOY.md

File metadata and controls

230 lines (171 loc) · 6.72 KB
Raw

Deployment Guide

Status: ✅ LIVE — System is deployed and operational.

Architecture

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   Frontend      │     │   Backend       │     │   Database      │
│   (Vercel)      │────▶│   (Render)      │────▶│ (InfluxDB Cloud)│
│   React + Vite  │     │   FastAPI       │     │   Time-Series   │
└─────────────────┘     └─────────────────┘     └─────────────────┘
Component Technology Hosting
Frontend React 18 + Vite Vercel
Backend FastAPI + Uvicorn + Docker Render
Database InfluxDB 2.x InfluxDB Cloud (AWS us-east-1)

🚀 Live Links

Service URL Status
Frontend https://predictive-maintenance-ten.vercel.app/ ✅ Live
Backend API https://predictive-maintenance-uhlb.onrender.com ✅ Live
API Docs https://predictive-maintenance-uhlb.onrender.com/docs ✅ Live
Health Check https://predictive-maintenance-uhlb.onrender.com/health ✅ Live

Deployment Instructions

Backend (Render)

Already deployed at https://predictive-maintenance-uhlb.onrender.com

Environment Variables (configured in Render Dashboard):

ENVIRONMENT=production
PORT=8000
INFLUX_URL=https://us-east-1-1.aws.cloud2.influxdata.com
INFLUX_TOKEN=<your-token>
INFLUX_ORG=<your-org-id>
INFLUX_BUCKET=sensor_data

Frontend (Vercel)

  1. Connect Repository:

    • Go to vercel.com
    • Import the GitHub repository
    • Set Root Directory to frontend

  2. Build Settings:

    • Framework: Vite
    • Build Command: npm run build
    • Output Directory: dist

  3. Deploy:

    • Click "Deploy"
    • Vercel will auto-configure rewrites from vercel.json

Local Development

Quick Start (Docker)

# Start all services
docker-compose up --build

# Access
# Frontend: http://localhost:5173
# Backend:  http://localhost:8000
# API Docs: http://localhost:8000/docs

Manual Setup

Backend:

cd backend
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt
uvicorn backend.api.main:app --reload

Frontend:

cd frontend
npm install
npm run dev

API Endpoints

Endpoint Method Description
/health GET Health check
/system/calibrate POST Start calibration (builds baseline + legacy + batch ML models)
/system/inject-fault POST Inject fault (type: spike/drift/jitter, severity: mild/medium/severe)
/system/reset POST Reset to idle
/system/purge POST Wipe all data + reset DI to 0.0
/system/status/{asset_id} GET Get system state (calibrating/monitoring/fault)
/api/v1/status/{asset_id} GET Get asset status (health, DI, RUL)
/api/v1/data/history/{asset_id} GET Get sensor history + events + health scoring
/api/v1/report/{asset_id} GET Generate PDF/Excel report (format: pdf/xlsx/industrial)
/sandbox/predict POST What-If analysis
/ping GET Keep-alive heartbeat (prevents Render cold starts)

Environment Variables Reference

Backend (Render Dashboard)

Variable Required Description Example
ENVIRONMENT Yes Runtime mode production
PORT Yes Server port 8000
INFLUX_URL Yes InfluxDB Cloud URL https://us-east-1-1.aws.cloud2.influxdata.com
INFLUX_TOKEN Yes InfluxDB API token kg2i8Mq...
INFLUX_ORG Yes InfluxDB Organization ID 67c4314d97304c09
INFLUX_BUCKET Yes InfluxDB bucket name sensor_data

Frontend (Vercel)

Variable Required Description Example
VITE_API_URL No Backend URL (handled via rewrites) https://predictive-maintenance-uhlb.onrender.com

Vercel Rewrites Configuration

The frontend/vercel.json file handles API proxying:

{
  "rewrites": [
    {
      "source": "/api/:path*",
      "destination": "https://predictive-maintenance-uhlb.onrender.com/api/:path*"
    },
    {
      "source": "/system/:path*",
      "destination": "https://predictive-maintenance-uhlb.onrender.com/system/:path*"
    },
    {
      "source": "/health",
      "destination": "https://predictive-maintenance-uhlb.onrender.com/health"
    },
    {
      "source": "/ping",
      "destination": "https://predictive-maintenance-uhlb.onrender.com/ping"
    }
  ]
}

This allows the frontend to call /api/... and have Vercel proxy to Render automatically.

Monitoring

  • InfluxDB Dashboard: https://cloud2.influxdata.com
  • Render Logs: Render Dashboard → Service → Logs
  • Vercel Logs: Vercel Dashboard → Project → Functions

⚠️ Common Pitfalls

Windows Pitfall: node_modules in Git

Problem: Committing node_modules/ from Windows uploads Windows-specific binaries (.cmd, .exe). When Vercel (Linux) tries to execute these, you get:

Error: Command "npm run build" exited with 126

Solution:

  1. Ensure node_modules/ is in .gitignore
  2. If already committed, remove with: 
    git rm -r --cached frontend/node_modules
git commit -m "fix: remove node_modules from tracking"
git push
  3. Redeploy on Vercel

Render Cold Starts

Free-tier Render services spin down after 15 minutes of inactivity. First request after idle may take 30-60 seconds.

The frontend sends a 10-minute keep-alive ping to /ping automatically to prevent this during active browser sessions.

Degradation Index (DI) State

The Cumulative Degradation Index is persisted to InfluxDB. On backend restart, DI is hydrated from InfluxDB via |> last(). The POST /system/purge endpoint writes DI=0.0 (InfluxDB Cloud Serverless v3 does not support range deletes).

InfluxDB Token Scope

Ensure your InfluxDB token has read/write access to the sensor_data bucket.

Troubleshooting

Backend not responding

curl https://predictive-maintenance-uhlb.onrender.com/health

Expected: {"status":"healthy","database":"connected",...}

Frontend API errors

  • Check browser DevTools → Network tab
  • Verify Vercel rewrites are working (requests to /api/* should proxy to Render)

InfluxDB connection issues

  • Verify credentials in Render environment variables
  • Check InfluxDB Cloud dashboard for bucket access