Project Interview Bible ⚡ Recently updated

Explaining Deployment in Your Project

Interviewers want to know your app runs in production — not just on localhost.

What to Cover

Where each part is deployed
How deploys work (manual vs CI/CD)
Environment variables and secrets
Database hosting
Domain and HTTPS

Typical Student Deployment Stack

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│  Frontend       │     │  Backend API    │     │  Database       │
│  Vercel         │────►│  Railway        │────►│  Railway        │
│  (or Netlify)   │     │  (or Render)    │     │  Postgres addon │
└─────────────────┘     └─────────────────┘     └─────────────────┘
     CDN included            Auto HTTPS              Managed backups

One-sentence summary: “React frontend on Vercel, Express API and PostgreSQL on Railway, all connected via environment variables, HTTPS everywhere.”

Deployment Script (What to Say)

“I deploy the frontend to Vercel — connected to my GitHub repo, so every push to main auto-deploys. Preview URLs are generated for every pull request.

The backend runs on Railway with a managed PostgreSQL database. Railway builds from my Dockerfile / package.json start script and sets PORT automatically.

Environment variables like DATABASE_URL, JWT_SECRET, and FRONTEND_URL are set in each platform’s dashboard — never committed to git.

The live app is at https://myapp.vercel.app and the API at https://myapi.railway.app.”

Environment Variables

Variable	Where	Purpose
`DATABASE_URL`	Backend	Postgres connection string
`JWT_SECRET`	Backend	Sign/verify tokens
`FRONTEND_URL`	Backend	CORS allowed origin
`VITE_API_URL` / `NEXT_PUBLIC_API_URL`	Frontend	API base URL
`NODE_ENV`	Backend	`production` enables optimizations

Security talking point: “.env is in .gitignore. I use .env.example with placeholder values so teammates know what’s needed. Secrets are in Railway/Vercel dashboard only.”

CI/CD (If You Have It)

Simple (Platform Auto-Deploy)

“Push to main → Vercel/Railway detects change → builds → deploys. Takes about 2 minutes.”

With GitHub Actions

# .github/workflows/deploy.yml (example)
name: Test and Deploy
on:
  push:
    branches: [main]
  pull_request:
    branches: [main]

jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
        with: { node-version: '20' }
      - run: npm ci
      - run: npm test
      # Vercel/Railway deploys automatically after merge if tests pass

What to say: “GitHub Actions runs tests on every PR. Merge to main triggers auto-deploy on Vercel and Railway. I don’t SSH into servers — it’s all git-push deploys.”

Database in Production

Topic	What to say
Hosting	”Railway managed Postgres — they handle backups and updates”
Migrations	”`prisma migrate deploy` runs on Railway’s build step”
Seeding	”Seed script for dev only — never run seed in production”
Backups	”Railway does daily backups. For critical data I’d also export weekly.”
Connection pooling	”Prisma handles connection pooling. At scale I’d add PgBouncer.”

CORS Configuration

// Only allow your frontend domain in production
const corsOptions = {
  origin: process.env.FRONTEND_URL, // https://myapp.vercel.app
  credentials: true, // if using cookies
};
app.use(cors(corsOptions));

Common bug story: “Initially my frontend couldn’t reach the API — browser blocked requests with CORS error. Fixed by setting FRONTEND_URL in Railway to match my Vercel domain exactly.”

Docker (If You Use It)

# Dockerfile (simplified)
FROM node:20-alpine
WORKDIR /app
COPY package*.json ./
RUN npm ci --only=production
COPY . .
EXPOSE 3000
CMD ["node", "src/server.js"]

What to say: “Dockerfile ensures the same Node version and dependencies in dev and production. Railway builds the image and runs it.”

Common Deployment Issues (Good Stories)

Issue	Story
Build fails on deploy	”Forgot to add `engines` in package.json — Railway used Node 16, I needed 20”
Env var missing	”API returned 500 — `JWT_SECRET` wasn’t set in Railway dashboard”
CORS errors	”Frontend on Vercel, API on Railway — had to whitelist exact Vercel URL”
Database connection	”Used `localhost` DATABASE_URL in production — switched to Railway’s provided URL”
Static files 404	”Forgot `VITE_API_URL` — frontend was calling localhost in production build”

Turn bugs into learning moments.

Common Interview Questions

Q: How do you deploy without downtime? A: “Railway does rolling deploys — new container starts, health check passes, old one stops. For a student app this is automatic. At scale I’d use blue-green deployment.”

Q: How do you handle different environments (dev/staging/prod)? A: “Local uses .env with local Postgres. Production uses Railway env vars. I could add a staging branch that deploys to a preview Railway environment.”

Q: What if the database goes down? A: “API returns 503 with a friendly message. Frontend shows ‘service unavailable’ instead of crashing. Railway auto-restarts the DB; I’d set up uptime monitoring with UptimeRobot.”

Q: How do you monitor production? A: “Railway dashboard shows CPU/memory logs. I added Morgan for request logging. For errors I’d add Sentry — it’s on my improvement list.”

Checklist

Know your live URL(s) by heart
Can explain frontend vs backend hosting
Can explain how env vars are managed
Can describe deploy flow (git push → auto deploy)
Have one deployment bug story ready

Next: performance.md | security.md