Deploying to Cloudflare Workers

Deploy your own ERA Agent instance on Cloudflare Workers to get a dedicated, customizable code execution platform running on the global edge network.

Why Deploy Your Own?

While you can use the public ERA Agent at https://anewera.dev, deploying your own instance gives you:

• 🔐 Private Access - Control who can use your instance
• ⚙️ Custom Configuration - Adjust limits, timeouts, and settings
• 📊 Usage Monitoring - Track your own metrics and logs
• 🎯 Custom Domain - Use your own domain name
• 💰 Cost Control - Optimize for your usage patterns
• 🔒 Data Privacy - Keep data in your Cloudflare account

Prerequisites

• Cloudflare Account - Sign up free
• Node.js 18+ - For running Wrangler CLI
• npm or yarn - Package manager
• Git - For cloning the repository

Quick Deployment

1. Clone the Repository

git clone https://github.com/yourusername/era-agent
cd era-agent/cloudflare

2. Install Dependencies

npm install

3. Configure Wrangler

Login to Cloudflare:

npx wrangler login

This will open a browser for you to authenticate.

4. Deploy

npx wrangler deploy

That’s it! Your ERA Agent is now deployed at:

https://your-worker-name.workers.dev

Configuration

Worker Settings

Edit wrangler.toml to customize your deployment:

name = "my-era-agent"
main = "src/index.ts"
compatibility_date = "2024-10-01"
node_compat = true

# Environment variables
[vars]
AGENT_LOG_LEVEL = "info"

# Durable Objects
[[durable_objects.bindings]]
name = "ERA_AGENT"
class_name = "EraAgent"

[[durable_objects.bindings]]
name = "SESSIONS"
class_name = "SessionDO"

# R2 Storage (for persistent files)
[[r2_buckets]]
binding = "SESSIONS_BUCKET"
bucket_name = "era-sessions"

# Container runtime bindings
[container_runtime]
# Configure container execution
instance_type = "standard"

Custom Domain

Use your own domain instead of *.workers.dev:

# Add route in wrangler.toml
[routes]
pattern = "era.yourdomain.com"
zone_name = "yourdomain.com"

# Deploy
npx wrangler deploy

Resource Limits

Adjust execution limits in src/index.ts:

// Default limits
const DEFAULT_TIMEOUT = 30; // seconds
const MAX_TIMEOUT = 300;    // 5 minutes
const DEFAULT_MEMORY = 256; // MiB
const MAX_MEMORY = 2048;    // 2 GB

Environment Variables

Set secrets for production:

# Set API key (optional)
npx wrangler secret put API_KEY

# Set custom limits
npx wrangler secret put MAX_EXECUTION_TIME

# Deploy with new secrets
npx wrangler deploy

Authentication

Add authentication to your deployment to restrict access.

Option 1: API Key Authentication

Add API key check in src/index.ts:

// Before handling requests
const apiKey = request.headers.get('Authorization');
const expectedKey = env.API_KEY; // Set via wrangler secret

if (apiKey !== `Bearer ${expectedKey}`) {
  return new Response('Unauthorized', { status: 401 });
}

Set the API key:

npx wrangler secret put API_KEY
# Enter your secret key when prompted

npx wrangler deploy

Use it:

curl -X POST https://your-worker.workers.dev/api/execute \
  -H "Authorization: Bearer your-secret-key" \
  -H "Content-Type: application/json" \
  -d '{"language":"python","code":"print(\"Hello!\")"}'

Option 2: Cloudflare Access

Protect your Worker with Cloudflare Access:

1. Go to Cloudflare Dashboard → Zero Trust → Access
2. Create new application
3. Add your Worker URL: https://your-worker.workers.dev
4. Configure authentication rules (email, SSO, etc.)
5. Users must authenticate before accessing

Option 3: IP Allowlist

Restrict access by IP address:

const ALLOWED_IPS = ['203.0.113.0', '198.51.100.0'];

if (!ALLOWED_IPS.includes(request.headers.get('CF-Connecting-IP'))) {
  return new Response('Forbidden', { status: 403 });
}

Storage Setup

R2 Bucket Creation

ERA Agent uses R2 for persistent file storage:

# Create R2 bucket
npx wrangler r2 bucket create era-sessions

# Verify
npx wrangler r2 bucket list

Update wrangler.toml:

[[r2_buckets]]
binding = "SESSIONS_BUCKET"
bucket_name = "era-sessions"

Durable Objects Setup

Durable Objects are automatically created. To manually manage:

# List Durable Objects
npx wrangler durable-objects list

# Get specific instance
npx wrangler durable-objects get <id>

Monitoring & Logs

Real-Time Logs

View live logs from your Worker:

# Tail logs
npx wrangler tail

# Filter logs
npx wrangler tail --format json

# Filter by status
npx wrangler tail --status error

Analytics Dashboard

View analytics in Cloudflare Dashboard:

1. Go to Workers & Pages → Your Worker
2. Click “Metrics” tab
3. View:
   • Requests per second
   • Success rate
   • CPU time
   • Duration metrics

Custom Logging

Add structured logging:

console.log(JSON.stringify({
  level: 'info',
  message: 'Code execution started',
  session_id: session.id,
  language: language,
  timestamp: new Date().toISOString()
}));

Performance Optimization

Caching

Enable caching for static resources:

// Cache responses for repeated requests
const cache = caches.default;
const cacheKey = new Request(url, request);
let response = await cache.match(cacheKey);

if (!response) {
  response = await handleRequest(request);
  ctx.waitUntil(cache.put(cacheKey, response.clone()));
}

Cold Start Optimization

Minimize cold start time:

// Preload commonly used modules
import { loadPythonRuntime } from './runtimes';

// Keep connections warm
setInterval(() => {
  // Ping Durable Object
}, 50000); // 50 seconds

Resource Limits

Optimize for your workload:

[container_runtime]
# Standard tier: 1 CPU, 256 MB
instance_type = "standard"

# Or upgrade for more resources
# instance_type = "standard-2"  # 2 CPU, 512 MB

Cost Management

Cloudflare Workers Pricing

Free Tier:

• 100,000 requests/day
• 10 ms CPU time per invocation

Paid Plan ($5/month):

• 10 million requests included
• Additional: $0.50/million requests
• CPU time: $0.02/million CPU ms

R2 Storage Pricing

Free Tier:

• 10 GB storage
• 1 million Class A operations/month
• 10 million Class B operations/month

Beyond Free:

• Storage: $0.015/GB/month
• Class A ops: $4.50/million
• Class B ops: $0.36/million

Optimizing Costs

// 1. Set reasonable timeouts
const MAX_TIMEOUT = 30; // Don't allow excessive CPU time

// 2. Clean up sessions automatically
const SESSION_TTL = 30 * 60 * 1000; // 30 minutes

// 3. Limit file storage
const MAX_FILE_SIZE = 10 * 1024 * 1024; // 10 MB

Production Checklist

Before going to production, ensure:

Security

• Add authentication (API keys, Cloudflare Access, or IP allowlist)
• Set rate limiting
• Configure CORS appropriately
• Rotate secrets regularly
• Enable HTTPS only

Configuration

• Set appropriate resource limits
• Configure timeout values
• Set up error monitoring
• Configure logging level
• Set up alerts

Storage

• Create R2 bucket for persistent files
• Configure Durable Objects
• Set up backup strategy
• Plan for data retention

Monitoring

• Enable Cloudflare Analytics
• Set up log aggregation
• Configure alerts for errors
• Monitor costs
• Track usage patterns

Deployment Strategies

Development → Staging → Production

Use Wrangler environments:

wrangler.toml

[env.staging]
name = "era-agent-staging"
vars = { AGENT_LOG_LEVEL = "debug" }

[env.production]
name = "era-agent-production"
vars = { AGENT_LOG_LEVEL = "warn" }

Deploy:

# Staging
npx wrangler deploy --env staging

# Production
npx wrangler deploy --env production

CI/CD with GitHub Actions

.github/workflows/deploy.yml

name: Deploy to Cloudflare

on:
  push:
    branches: [main]

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - uses: actions/setup-node@v3
        with:
          node-version: '18'

      - name: Install dependencies
        run: npm install
        working-directory: ./cloudflare

      - name: Deploy to Cloudflare
        run: npx wrangler deploy
        working-directory: ./cloudflare
        env:
          CLOUDFLARE_API_TOKEN: ${{ secrets.CLOUDFLARE_API_TOKEN }}

Blue-Green Deployments

# Deploy new version to staging
npx wrangler deploy --env staging

# Test thoroughly
curl https://era-agent-staging.workers.dev/health

# Promote to production
npx wrangler deploy --env production

Troubleshooting

”Exceeded CPU time”

Problem: Code execution taking too long

Solutions:

• Reduce MAX_TIMEOUT setting
• Upgrade to higher tier Worker
• Optimize code execution
• Add execution time monitoring

”Durable Object not found”

Problem: Durable Object not initialized

Solutions:

# Redeploy to initialize
npx wrangler deploy

# Check Durable Object bindings
npx wrangler durable-objects list

“R2 bucket not accessible”

Problem: R2 bucket not created or bound

Solutions:

# Create bucket
npx wrangler r2 bucket create era-sessions

# Verify binding in wrangler.toml
[[r2_buckets]]
binding = "SESSIONS_BUCKET"
bucket_name = "era-sessions"

# Redeploy
npx wrangler deploy

Deployment Errors

Check Wrangler version:

npx wrangler --version
npm install -g wrangler@latest

Check authentication:

npx wrangler whoami
npx wrangler login

Clear cache:

rm -rf node_modules
npm install
npx wrangler deploy

Next Steps

• API Reference - Complete API documentation
• MCP Remote Server - Use as remote MCP
• Session Management - Advanced session features
• Cloudflare Docs - Workers documentation

Support

• 📚 ERA Agent Docs
• 💬 GitHub Issues
• 📖 Cloudflare Workers Docs
• 🎓 Cloudflare Community