How would you design a large-scale ticketing system that can handle millions of users during peak events (e.g., Black Friday)?

To handle millions of users during peak events, a large-scale ticketing system requires a robust architecture. I'd use a microservices approach to ensure independent scaling of components. Load balancing is critical to distribute traffic effectively. Caching frequently accessed data will reduce database load. For data storage, a hybrid approach using NoSQL for high-volume, less structured data and relational databases for transactional integrity (like bookings and payments) would be optimal. Key functionalities include real-time seat selection, secure payment processing, mobile-first design with QR code validation for entry, and comprehensive admin dashboards for monitoring and analytics. Prioritizing availability and fault tolerance through redundancy and graceful degradation is paramount.