The Hosting Landscape

Every Node.js application needs somewhere to run. The right choice depends on your traffic, budget, team expertise, and operational needs.

                               Hosting Options
                                     │
            ┌────────────────────────┼────────────────────────┐
            │                        │                        │
           VPS                   PaaS                   Serverless
            │                        │                        │
  ┌─────────┼─────────┐    ┌─────────┼─────────┐    ┌─────────┼─────────┐
  │         │         │    │         │         │    │         │         │
DigitalOcean AWS EC2    Heroku    Railway    Netlify  AWS Lambda CloudFlare
Linode     GCP Compute  Render    Fly.io     Vercel   GCP Functions Workers

1. VPS (Virtual Private Server)

You get a virtual machine with root access. You manage everything - OS, Node.js, reverse proxy, SSL.

Provider	Starting Price	Specs
DigitalOcean	$6/mo	1 CPU, 1GB RAM, 25GB SSD
Linode	$5/mo	1 CPU, 1GB RAM, 25GB SSD
Vultr	$6/mo	1 CPU, 1GB RAM, 25GB SSD
AWS EC2 (t2.micro)	Free tier (1yr)	1 CPU, 1GB RAM
Hetzner	$4/mo	2 CPU, 2GB RAM, 40GB SSD

Setting Up a Node.js App on VPS

# 1. SSH into your server
ssh root@your-server-ip

# 2. Install Node.js
curl -fsSL https://deb.nodesource.com/setup_22.x | bash -
apt-get install -y nodejs nginx

# 3. Clone your app
git clone https://github.com/your/repo.git /var/www/app
cd /var/www/app
npm ci --production

# 4. Set up environment variables
cat > /etc/environment << EOF
NODE_ENV=production
PORT=3000
DATABASE_URL=postgres://...
EOF

# 5. Start with PM2
npm install -g pm2
pm2 start app.js --name my-api
pm2 save
pm2 startup  # Auto-start on reboot

# 6. Configure Nginx as reverse proxy
cat > /etc/nginx/sites-available/my-api << EOF
server {
    listen 80;
    server_name your-domain.com;

    location / {
        proxy_pass http://localhost:3000;
        proxy_http_version 1.1;
        proxy_set_header Upgrade \$http_upgrade;
        proxy_set_header Connection 'upgrade';
        proxy_set_header Host \$host;
        proxy_set_header X-Real-IP \$remote_addr;
        proxy_cache_bypass \$http_upgrade;
    }
}
EOF

# 7. Enable HTTPS with Certbot
apt-get install -y certbot python3-certbot-nginx
certbot --nginx -d your-domain.com

When to Use VPS

• You need full control over the environment
• Traffic is predictable and consistently moderate-to-high
• You have DevOps experience or want to learn
• Budget is tight - VPS offers the best performance per dollar
• You need specific OS-level packages or configurations

VPS Pros & Cons

Pros	Cons
Complete control	You manage everything (OS updates, security)
Best performance per dollar	Manual scaling (add more VPS, load balancer)
No vendor lock-in	SSH and command-line required
Predictable pricing	No built-in monitoring or logging

2. PaaS (Platform as a Service)

PaaS providers manage the infrastructure. You just push code.

Provider	Starting Price	Features
Railway	$5/mo	Quick deploy, built-in Postgres, auto HTTPS
Render	$7/mo	Auto deploy from Git, managed Postgres/Redis
Fly.io	Free tier	Edge compute, global regions, WireGuard VPN
Heroku	$7/mo (eco)	Mature ecosystem, add-ons marketplace
DigitalOcean App Platform	$5/mo	Integrated with DO VPS ecosystem

Deploying to Railway (Example)

# 1. Install Railway CLI
npm install -g @railway/cli

# 2. Login and init
railway login
railway init

# 3. Set environment variables
railway variables set NODE_ENV=production
railway variables set DATABASE_URL=...

# 4. Deploy
railway up

# 5. Open the deployed app
railway open

Railway railway.json

{
  "build": {
    "builder": "nixpacks",
    "buildCommand": "npm ci --production"
  },
  "deploy": {
    "startCommand": "node app.js",
    "healthcheckPath": "/health",
    "restartPolicyType": "always"
  }
}

When to Use PaaS

• You want to focus on code, not infrastructure
• Traffic is low-to-moderate and you don’t want to over-provision
• You want zero-downtime deploys out of the box
• You need managed databases (Postgres, Redis) without admin work

PaaS Pros & Cons

Pros	Cons
Minimal DevOps knowledge needed	More expensive than VPS at scale
Auto-scaling (limited)	Vendor lock-in
Built-in logging and monitoring	Limited control over environment
HTTPS, CDN, domains managed for you	Cold starts (sleep after inactivity)

3. Serverless

Your code runs in stateless containers that scale to zero. You pay per invocation.

Provider	Free Tier	Limits
AWS Lambda	1M requests/mo	15min timeout, 10GB RAM
Google Cloud Functions	2M invocations/mo	60min timeout, 16GB RAM
Netlify Functions	125K requests/mo	10s timeout, 1GB RAM
Vercel Functions	100GB-hours/mo	60s timeout (Hobby)
Cloudflare Workers	100K requests/day	50ms CPU time (paid up to 30s)

Serverless Node.js Handler

// AWS Lambda handler
exports.handler = async (event) => {
  try {
    const users = await db.query('SELECT * FROM users LIMIT 10');

    return {
      statusCode: 200,
      headers: { 'Content-Type': 'application/json' },
      body: JSON.stringify(users),
    };
  } catch (err) {
    return {
      statusCode: 500,
      body: JSON.stringify({ error: err.message }),
    };
  }
};

When to Use Serverless

• Sporadic traffic - a few requests per hour, or bursty
• Background jobs - image processing, email sending, webhooks
• Prototyping - no cost when idle
• Microservices - each function is a small, independent unit

Serverless Pros & Cons

Pros	Cons
Pay per request (cheap at low volume)	Cold starts (100ms–1s delay)
Auto-scales infinitely	15-minute timeout (AWS Lambda)
No server management	Stateless - no local filesystem
Built-in fault tolerance	Debugging is harder (no SSH)

4. Comparison Matrix

Feature	VPS	PaaS	Serverless
Setup time	Hours	Minutes	Minutes
DevOps skill needed	High	Low	Medium
Monthly cost (low traffic)	$5–10	$5–7	$0 (free tier)
Monthly cost (10K req/s)	$50–200	$200–1000+	$500+
Cold start	None	1-5s (if sleep)	100ms–1s
Max request duration	Unlimited	Unlimited (with config)	15min (Lambda)
Database options	Any	Managed (limited)	External only
Custom domain + SSL	Manual	Auto	Auto
Monitoring	DIY	Built-in	Built-in

5. Docker-Based Deployment

A Docker container can run on any of these platforms:

FROM node:22-alpine

WORKDIR /app
COPY package*.json ./
RUN npm ci --production
COPY . .

EXPOSE 3000
CMD ["node", "app.js"]

# Run on VPS
docker build -t my-api .
docker run -d -p 3000:3000 --name my-api my-api

# Deploy to Railway (auto-detect Dockerfile)
railway up

# Deploy to AWS ECS / Google Cloud Run
gcloud run deploy my-api --source .

Making the Choice

Ask these questions to decide:

1. How much traffic? Low → PaaS/Serverless. High → VPS.
2. Do you want to manage servers? No → PaaS or Serverless. Yes → VPS.
3. Budget? Lowest → VPS. Predictable → PaaS. Usage-based → Serverless.
4. Need databases? PaaS offers managed ones. VPS needs manual setup.
5. Team size? Small team → PaaS (less ops overhead). Large team → VPS (more control).
6. Scaling pattern? Steady → VPS. Bursty → Serverless. Growing → PaaS.

Key Takeaways

• VPS (DigitalOcean, Linode) - best performance per dollar, full control, highest ops overhead
• PaaS (Railway, Render, Heroku) - easiest deployment, managed infrastructure, moderate cost
• Serverless (AWS Lambda, Vercel) - pay per request, scales to zero, cold starts are the trade-off
• No perfect choice - it depends on traffic patterns, team skills, budget, and operational needs
• Start simple - deploy to PaaS first (minutes), migrate to VPS if cost becomes an issue
• Use Docker to avoid vendor lock-in - a Docker image runs on any platform
• Vendor lock-in is real - PaaS-specific features (queues, databases) are hard to migrate away from