TL;DR — Cascading Failure in Microservices

You operate a ride-hailing platform with 12 microservices. The architecture: Mobile App -> API Gateway -> Ride Service -> Payment Service, Pricing Service, Driver Service, Notification Service. Each service calls downstream services synchronously with a 30-second default timeout. The system handles 50K ride requests per minute normally. Key features: Payment failures must not block ride requests — allow ride now, charge later. Each service failure must be isolated to its own domain.