# Server Recommendation System - Setup Guide

## 1. Create Cloudflare Resources

### Create D1 Database
```bash
# Create D1 database
npx wrangler d1 create server-recommend-db

# Copy the database_id from output and update wrangler.toml
```

### Create KV Namespace
```bash
# Create KV namespace for caching
npx wrangler kv:namespace create CACHE

# Copy the id from output and update wrangler.toml
```

## 2. Initialize Database Schema

```bash
# Execute schema.sql against your D1 database
npx wrangler d1 execute server-recommend-db --file=schema.sql
```

## 3. Seed Sample Data (Optional)

Create a file `seed.sql` with sample server data and run:
```bash
npx wrangler d1 execute server-recommend-db --file=seed.sql
```

## 4. Development

```bash
# Start local development server
npm run dev

# Type checking
npm run typecheck
```

## 5. Deployment

```bash
# Deploy to Cloudflare Workers
npm run deploy
```

## API Endpoints

### Health Check
```bash
curl https://your-worker.workers.dev/api/health
```

### Get Servers
```bash
curl "https://your-worker.workers.dev/api/servers?minCpu=4&minMemory=8"
```

### Request Recommendations
```bash
curl -X POST https://your-worker.workers.dev/api/recommend \
  -H "Content-Type: application/json" \
  -d '{
    "cpu_cores": 4,
    "memory_gb": 8,
    "storage_gb": 100,
    "network_bandwidth_mbps": 1000,
    "sla_requirement": 99.9,
    "budget_monthly": 200,
    "regions": ["us-east-1", "us-west-2"]
  }'
```

## Environment Variables

No environment variables needed - all bindings are configured in `wrangler.toml`:
- `AI`: Workers AI binding (automatic)
- `DB`: D1 Database binding
- `CACHE`: KV namespace binding

## Testing Workers AI Integration

The worker uses `@cf/meta/llama-3.1-8b-instruct` model. Test with:

```bash
# Local development automatically uses Workers AI
npm run dev

# Make test request
curl -X POST http://localhost:8787/api/recommend \
  -H "Content-Type: application/json" \
  -d '{
    "cpu_cores": 2,
    "memory_gb": 4,
    "storage_gb": 50
  }'
```

## Cache Strategy

Recommendations are cached in KV for 1 hour based on request parameters:
- Cache key format: `recommend:cpu:X|mem:Y|stor:Z|...`
- TTL: 3600 seconds (1 hour)
- Cache hit/miss logged in console

## Error Handling

All endpoints return proper HTTP status codes:
- `200`: Success
- `400`: Invalid request (validation error)
- `404`: Endpoint not found
- `500`: Internal server error

CORS is enabled for all origins (`Access-Control-Allow-Origin: *`).

## Monitoring

Enable observability in `wrangler.toml`:
```toml
[observability]
enabled = true
```

View logs:
```bash
npx wrangler tail
```