At tiket.com, managing billions of hotel inventory rows pushed PostgreSQL to its limits, with write delays of up to four hours. After testing alternatives, we migrated to ScyllaDB and saw dramatic gains: 200x higher write throughput, 12x faster reads, and 60% lower costs. This talk shares our migration journey, the architectural reasons behind ScyllaDB’s performance, and the trade-offs we faced adapting our workloads to its stricter query model.