
As transaction volume grows in products like Revolut Card payments, bank transfers, and ledger updates, a single PostgreSQL instance can become a bottleneck on writes. Interviewers ask this to test whether you can separate techniques that improve read scale, write scale, and operational manageability.
Explain how replication, sharding, and partitioning help scale a database under heavy write loads. You should compare what each technique actually improves, where it does not help, and what trade-offs it introduces around consistency, routing, rebalancing, and operational complexity.
A strong answer should distinguish vertical vs horizontal scaling, explain why replication alone usually does not solve a primary write bottleneck in PostgreSQL, show how partitioning can reduce index and maintenance overhead on large write-heavy tables, and describe when sharding is the real write-scaling step. You should also mention implications for transactional correctness and data integrity in financial systems.