Hugging Face's logo

Spaces:
Pulastya0
/
Data-Science-Agent
Running

App Files Community
Data-Science-Agent / RENDER_DEPLOYMENT.md
Pulastya B
feat: Add Render deployment configuration
c10a976
preview code
|
raw
history blame contribute delete
6.79 kB

Render Deployment Guide

Prerequisites

  1. A Render account (free tier available)
  2. Your GitHub repository connected to Render
  3. Google Gemini API key

Quick Deploy (Recommended)

Option 1: Using render.yaml (Infrastructure as Code)

  1. Push your code to GitHub (already done)

  2. Create a new Web Service on Render:

    • Go to https://dashboard.render.com/
    • Click "New +" → "Blueprint"
    • Connect your GitHub repository: Pulastya-B/DevSprint-Data-Science-Agent
    • Render will automatically detect the render.yaml file
    • Click "Apply"

  3. Add Secret Environment Variable:

    • Go to your service dashboard
    • Navigate to "Environment" tab
    • Add your GOOGLE_API_KEY (this is sensitive and not included in render.yaml)
    • Click "Save Changes"

  4. Deploy:

    • Render will automatically build and deploy your application
    • Wait for the build to complete (~5-10 minutes for first deploy)
    • Your app will be available at: https://data-science-agent.onrender.com

Option 2: Manual Setup

  1. Create a new Web Service:

  2. Configure the service:

    • Name: data-science-agent
    • Region: Oregon (US West)
    • Branch: main
    • Runtime: Docker
    • Plan: Free (or Starter for production)

  3. Add Environment Variables:

    LLM_PROVIDER=gemini
GOOGLE_API_KEY=<your-api-key-here>
GEMINI_MODEL=gemini-2.5-flash
REASONING_EFFORT=medium
CACHE_DB_PATH=/tmp/cache_db/cache.db
CACHE_TTL_SECONDS=86400
OUTPUT_DIR=/tmp/outputs
DATA_DIR=/tmp/data
MAX_PARALLEL_TOOLS=5
MAX_RETRIES=3
TIMEOUT_SECONDS=300
PORT=8080
ARTIFACT_BACKEND=local

  4. Configure Health Check:

    • Health Check Path: /api/health

  5. Deploy:

    • Click "Create Web Service"
    • Wait for the build to complete

Important Notes

Free Tier Limitations

  • Spin down after inactivity: Free tier services spin down after 15 minutes of inactivity
  • Cold starts: First request after spin-down will take 30-60 seconds
  • Memory: 512 MB RAM (may be tight for large ML models)
  • Build time: Free tier has slower build times

Upgrading to Paid Plan

For production use, consider upgrading to at least the Starter plan ($7/month):

  • No spin-down
  • Faster builds
  • More memory (512 MB → 2 GB)
  • Better performance

Storage Considerations

  • Render uses ephemeral storage - files are lost on restart
  • For persistent storage, consider:
    • Connecting to external storage (S3, GCS)
    • Using Render's persistent disk (paid plans only)
    • Storing only temporary analysis results

Performance Optimization

  1. Use caching: The app includes SQLite caching for repeated queries
  2. Monitor memory usage: Large datasets may exceed free tier limits
  3. Optimize docker image: The multi-stage build already optimizes image size
  4. Regional selection: Choose a region close to your users

Deployment Commands

Manual Rebuild (if needed) 

# Trigger rebuild via Render Dashboard
# or use Render API
curl -X POST https://api.render.com/v1/services/<service-id>/deploys \
  -H "Authorization: Bearer <your-api-key>"

Check Logs 

# View logs in Render Dashboard
# or use Render CLI
render logs -s data-science-agent

Custom Domain (Optional)

  1. Go to your service dashboard
  2. Click "Settings" → "Custom Domain"
  3. Add your domain (e.g., agent.yourdomain.com)
  4. Update your DNS records as instructed
  5. Render automatically provisions SSL certificates

Troubleshooting

Build Fails

Issue: Docker build timeout

  • Solution: Increase build timeout in Render settings
  • Alternative: Optimize Dockerfile to reduce build time

Issue: Out of memory during build

  • Solution: Upgrade to paid plan with more memory
  • Alternative: Reduce dependencies in requirements.txt

App Crashes on Startup

Issue: Missing environment variables

  • Solution: Verify all required env vars are set in Render dashboard

Issue: Port binding error

  • Solution: Ensure app listens on 0.0.0.0 and PORT env variable

Slow Performance

Issue: Cold starts on free tier

  • Solution: Upgrade to paid plan to prevent spin-down
  • Workaround: Use a cron job to ping your app every 10 minutes

Issue: Large dataset processing timeout

  • Solution: Increase TIMEOUT_SECONDS env variable
  • Consider: Processing large datasets asynchronously

Monitoring

Health Check

Your app exposes a health check endpoint at /api/health:

curl https://data-science-agent.onrender.com/api/health

Logs

  • View real-time logs in Render Dashboard
  • Configure log drains for external monitoring (paid plans)

Metrics

Render provides built-in metrics:

  • CPU usage
  • Memory usage
  • Request count
  • Response time

Security Best Practices

  1. Never commit API keys to Git (use environment variables)
  2. Enable CORS only for trusted domains in production
  3. Use HTTPS (Render provides this automatically)
  4. Rotate API keys regularly
  5. Monitor usage to detect anomalies

Cost Estimation

Free Tier

  • Cost: $0/month
  • Best for: Development, testing, hackathons
  • Limitations: Spin-down, slower builds, 512MB RAM

Starter Plan ($7/month)

  • No spin-down
  • 512MB RAM → 2GB RAM
  • Faster builds
  • Better for: Small production apps

Standard Plan ($25/month)

  • 4GB RAM
  • High performance
  • Best for: Production apps with moderate traffic

Deployment Checklist

  • Code pushed to GitHub
  • render.yaml committed to repository
  • Render account created
  • GitHub repository connected to Render
  • Blueprint deployed (or manual service created)
  • GOOGLE_API_KEY added as secret environment variable
  • Health check endpoint verified
  • Application accessible at Render URL
  • Custom domain configured (optional)
  • Monitoring and alerts set up

Support

Next Steps

After successful deployment:

  1. Test the deployment:

    curl https://data-science-agent.onrender.com/api/health

  2. Upload a test dataset via the web interface

  3. Monitor logs for any errors

  4. Configure custom domain (optional)

  5. Set up monitoring and alerts

  6. Share your deployed app! 🚀

Your app will be live at: https://data-science-agent.onrender.com

(URL will be different if you choose a different service name)