A Kubernetes-native train ticketing platform designed to handle extreme traffic surges during peak seasons (Eid). Solves critical issues of the current system including crashes, double-booking, and failed transactions under 30M+ requests.
Current Bangladesh Railway online ticketing (Shohoz-Synesis JV) faces:
- 1,187+ booking attempts per seat during Eid
- 30M hits in 30 minutes causing system crashes
- Frequent OTP & payment failures
- Double-booking and unfair ticket allocation
- Atomic Seat Reservation β Redis-based locks with TTL prevent double-booking
- Auto-Scaling β Kubernetes HPA handles traffic spikes automatically
- Distributed Tracing β OpenTelemetry & Jaeger for full request observability
- Multi-Platform β Web (Next.js), Mobile (React Native), Desktop (Electron)
- Event-Driven β Kafka/SQS for async booking & notifications
- Zero Downtime β Rolling deployments with health checks