How do you design a system to handle cascading failures in microservices?

Cascading failures are prevented using circuit breakers, bulkheads, timeouts, and rate limiting.

When one service fails, dependent services may get overloaded. Circuit breakers stop requests to failing services. Bulkheads isolate resources. Timeouts prevent thread blocking. Load shedding drops excess traffic under pressure.