I've worked with microservices at scale and it was fantastic. We couldn't break backwards compatibility with our API without a lot of coordination. Outside of that, you could deploy as frequently as needed and other services could update as needed to make use of new features.

The monoliths I have worked in, very contrastingly, have had issues coordinating changes within the codebases, code crosses boundaries it should not and datastores get shared and coupled to (what should be) different domains leading to slow, inefficient code and ossified options for product changes.