when to break up a monolith into microservices?

hey everyone, ive been thinking a lot about when the right time to start splitting a monolith into microservices really is. weve got a growing application and while its been serving us well, im starting to see some pains with a single codebase. but, i also know that microservices introduce their own set of complexities. what were your tipping points? did you wait until you hit performance bottlenecks, or was it more about team structure and deployment velocity? looking for real-world experiences and common early warning signs. thanks!