Database Sharding: Strategies for Scaling Beyond Read Replicas

we've hit a scaling wall with our current database setup. Read replicas are helping, but the write load is becoming a bottleneck. We're starting to look into database sharding as the next step. has anyone here implemented sharding? What strategies did you find most effective, like range-based, hash-based, or directory-based? What were the biggest challenges or unexpected issues you ran into, especially around re-sharding or maintaining data consistency?