Challenge — Production Deployment

Objective

Deploy a full-stack app to production: NestJS API on Railway, Next.js frontend on Vercel, PostgreSQL on Supabase — with monitoring, CI/CD, and Docker.

Requirements

Infrastructure (50 points)

• Supabase project created with production database
• Prisma schema migrated to Supabase (migrate deploy)
• NestJS API deployed to Railway via Docker
• Next.js frontend deployed to Vercel
• Frontend can make requests to the Railway API (CORS configured)
• Both services using Supabase as their database

Quality (30 points)

• Environment variables set in Railway and Vercel dashboards (not in code)
• env.ts Zod validation — app crashes on startup if vars missing
• Health check endpoint at GET /health returning { status: 'ok' }
• Railway health check configured pointing to /health
• Uptime monitor watching the health endpoint

Monitoring (20 points)

• Sentry configured for both frontend and backend
• Structured logging with Pino (or NestJS Logger) in the API
• Vercel Analytics on the frontend
• Demonstrate: trigger an error, verify it appears in Sentry

Deployment Architecture

         Users
           |
    +------+------+
    |             |
Vercel (CDN)  Railway API
Next.js App   NestJS + Docker
vercel.app    railway.app
                  |
             Supabase
             PostgreSQL

Grading

Criteria	Points
Supabase database live with migrations	10
NestJS live on Railway (accessible URL)	15
Next.js live on Vercel (accessible URL)	15
API responding to requests from frontend	10
Health check endpoint working	5
Env vars in platform dashboards	10
Zod env validation	5
Sentry configured (both ends)	10
Uptime monitor active	5
CI/CD triggers on push to main	15
Total	100

Submission

Submit:

1. Frontend URL (Vercel)
2. API URL (Railway)
3. Screenshot of Railway dashboard showing your service running
4. Screenshot of Vercel dashboard showing your deployment
5. Screenshot of Sentry showing at least one event captured
6. Link to your GitHub repo with CI/CD workflows

Bonus

• Blue/green deployments (zero downtime)
• Separate staging environment (auto-deploy on PR)
• Database read replica on Supabase
• CDN for static assets (Cloudflare R2)
• Custom domain with SSL for both services