Skip to content

DEPLOYMENT.md

Latest commit

 

History

History
217 lines (149 loc) · 4.12 KB

DEPLOYMENT.md

File metadata and controls

217 lines (149 loc) · 4.12 KB

Deployment Guide

This document provides step-by-step instructions for deploying the MCP Server to various platforms.

Railway Deployment (Recommended)

Railway provides the easiest deployment experience for this MCP server.

Step 1: Prepare Your Repository

  1. Ensure all files are committed to Git
  2. Push to GitHub if not already there

Step 2: Deploy to Railway

  1. Visit railway.app and sign up/login
  2. Click "New Project"
  3. Select "Deploy from GitHub repo"
  4. Choose this repository
  5. Railway will automatically:
    • Detect it's a Node.js project
    • Install dependencies with npm install
    • Build the project with npm run build
    • Start the server with npm start

Step 3: Configure Environment (Optional)

Railway automatically provides:

  • PORT environment variable
  • Domain/URL for your deployment

No additional configuration needed!

Step 4: Access Your MCP Server

After deployment, you'll get a URL like:

Your MCP endpoint will be:

https://your-url.railway.app/mcp

Heroku Deployment

Prerequisites

  • Heroku CLI installed
  • Heroku account

Steps

  1. Create a Heroku app:
heroku create your-app-name
  1. Set buildpack (optional, auto-detected):
heroku buildpacks:set heroku/nodejs
  1. Deploy:
git push heroku main
  1. The app will be available at:
https://your-app-name.herokuapp.com/mcp

Vercel Deployment

Prerequisites

  • Vercel CLI or GitHub integration

Option 1: Vercel CLI

  1. Install Vercel CLI:
npm i -g vercel
  1. Deploy:
vercel

Option 2: GitHub Integration

  1. Visit vercel.com
  2. Import your GitHub repository
  3. Vercel will auto-deploy

Note: Vercel works best with serverless functions. This Express server works but may have cold start delays.

Docker Deployment

Dockerfile

Create a Dockerfile:

FROM node:18-alpine

WORKDIR /app

COPY package*.json ./
RUN npm ci --only=production

COPY . .
RUN npm run build

EXPOSE 3000

CMD ["npm", "start"]

Build and Run

docker build -t starbucks-mcp .
docker run -p 3000:3000 starbucks-mcp

Environment Variables

The server supports these environment variables:

Variable Description Default
PORT Port to listen on 3000
NODE_ENV Environment mode production

Health Checks

All platforms can use the built-in health check:

GET /health

Response:

{
  "status": "healthy",
  "service": "starbucks-mcp-server",
  "version": "1.0.0",
  "timestamp": "2023-..."
}

Monitoring

Basic Monitoring

The server logs to stdout/stderr, which most platforms capture automatically.

Advanced Monitoring

For production deployments, consider:

  1. Uptime monitoring: Ping /health endpoint
  2. Error tracking: Services like Sentry
  3. Performance monitoring: New Relic, DataDog, etc.
  4. Log aggregation: Logflare, Papertrail, etc.

Scaling

Railway

  • Automatic scaling available in pro plans
  • Handles traffic spikes automatically

Heroku

  • Use dynos for scaling:
heroku ps:scale web=2

Other Platforms

  • Most platforms offer horizontal scaling
  • This server is stateless and scales well

Troubleshooting

Common Issues

  1. Port binding errors: Ensure PORT environment variable is set
  2. Build failures: Check Node.js version (requires 18+)
  3. Memory issues: Most platforms provide 512MB+ which is sufficient

Debugging

  1. Check application logs
  2. Test /health endpoint
  3. Verify MCP endpoint with curl:
curl -X GET https://your-url/mcp \
  -H "Accept: text/event-stream"

Getting Help

  • Check platform-specific documentation
  • Use platform support channels
  • Open an issue in this repository

Security

HTTPS

All production deployments should use HTTPS. Most platforms provide this automatically.

CORS

The server has CORS enabled by default for development. Consider restricting origins in production.

Rate Limiting

Consider adding rate limiting for production deployments using middleware like express-rate-limit.