Multi-region database replication: our journey from single to global

we're a Series-C startup that's been primarily US-centric, but with increasing international customers, we're facing latency issues, particularly for users in Europe and Asia. a user in Singapore might experience 250ms+ latency accessing our database backend in the US, which is unacceptable for real-time interactions. we need a multi-region database strategy. we currently use PostgreSQL with read replicas, but replication lag and write latency across regions are big challenges. we're evaluating options like globally distributed databases (CockroachDB, YugabyteDB), active-active multi-master setups, or leveraging edge functions with local caching. data residency requirements for certain regions also complicate things. for those who've gone from a single-region to a global database setup, what was your journey like? what were the biggest hurdles and the solutions you found most effective?