A high-concurrency ticket booking engine built to solve the race condition problem: when thousands of users click 'Buy' simultaneously, how do you prevent double-bookings?

I implemented atomic distributed locking using Redis Lua scripts with SET NX, ensuring hold and release operations are race-condition free. Redis Pipelining batches thousands of seat state fetches into a single network roundtrip.

Real-time seat map updates are pushed instantly to all connected clients via WebSockets — when one user holds a seat, every other user sees it go grey immediately. The backend is built in Go with Fiber for high-performance HTTP and WebSocket handling, with PostgreSQL for ACID-compliant transaction recording.

• GitHubhttps://github.com/Manpreet-Bhatti/Ticketmain
• PlatformWeb
• StackGo, Fiber, Redis, PostgreSQL, React, TypeScript, Docker