What design patterns would you implement to handle partial failures and cascading latency in a microservices ecosystem?
