When should we actually think about using more than one database type for our microservices?

Hey everyone, I've been thinking about how we handle databases for microservices. You always hear that each microservice should have its own data, but a lot of teams, including us sometimes, just start with one database, maybe a relational one, that all the services talk to. It's way simpler at first. But when does that actually become a problem or a bad idea? I'm really interested in what you all have seen. Like, what were the big signs that told you it was time to stop using one big database and start giving different services their own data stores? Was it things getting slow, schemas being too tangled, deployments becoming a headache, or something else entirely?