Design a photo viewing application similar to Google Photos that supports millions of users uploading, storing, viewing, and searching billions of photos. The system must handle high-throughput uploads (10k+ photos/sec globally), store petabytes of image data reliably, and serve photos to users with low latency worldwide. Users should be able to upload photos, organize them into albums, apply tags, search by content/metadata, and browse through feeds. The system should automatically generate multiple resolutions (thumbnail, medium, full) for each photo. Implement features like face detection for people grouping, location-based search, and time-based browsing. Ensure the system is highly available (99.9% uptime), scales horizontally, and provides real-time monitoring and alerting for operations teams.