DO GPU Notifier

Real-time GPU availability monitoring for DigitalOcean with push notifications. React frontend + Node.js backend.

Features

• Real-time monitoring of DigitalOcean GPU availability
• Push notifications when GPUs become available
• Support for L40S, H100-1X, and H100-8X GPU instances
• Configurable regions and user-friendly interface

Quick Start

1. Clone and setup:

   git clone <repository-url>
   cd gpu-cap-ui

2. Backend setup:

   cd backend
   npm install

3. Create .env file:

   PORT=3001
   DO_API_TOKEN=your_digitalocean_api_token
   VAPID_PUBLIC_KEY=your_vapid_public_key
   VAPID_PRIVATE_KEY=your_vapid_private_key
   VAPID_SUBJECT=mailto:your-email@example.com
   REGIONS_TO_CHECK=nyc1,sfo3,fra1

4. Generate VAPID keys:

   npx web-push generate-vapid-keys

5. Frontend setup:

   cd ../frontend
   npm install

6. Run the application:

   # Terminal 1 - Backend
   cd backend && npm run dev
   
   # Terminal 2 - Frontend
   cd frontend && npm start

Configuration

Variable	Description	Default
DO_API_TOKEN	DigitalOcean API token	Required
VAPID_PUBLIC_KEY	VAPID public key	Required
VAPID_PRIVATE_KEY	VAPID private key	Required
REGIONS_TO_CHECK	Regions to monitor	nyc1,sfo3,fra1

API Endpoints

• GET /api/health - Health check
• GET /api/status - Current GPU availability
• POST /api/status/check - Manual availability check
• POST /api/subscription - Subscribe to notifications
• GET /api/subscription/:userId - Get subscription info

How It Works

1. Backend scheduler checks GPU availability every 100 seconds
2. Uses DigitalOcean's capacity API to monitor regions
3. Sends push notifications when GPUs become available
4. Frontend provides interface for managing preferences

Troubleshooting

• API Errors: Verify DigitalOcean API token permissions
• Push Notifications: Check VAPID key configuration
• Rate Limiting: System includes delays between API calls

License

MIT License