Gaining complexity in a microservices system certainly isn't for the faint of heart (though neither is complexity in monoliths!). When there are many services that need to communicate with one another, we might need to coordinate multiple services communicating with multiple other services. We also might code for varying environments such as local, development server, or the cloud.
How do services know where to find one another? How can we avoid problems when a service is unavailable? How do we handle requests when we scale up or down certain parts of our system?