FabHotels interview question

When do you get less latency, accessing services through microservices or using monoliths?