The article explores the limitations of the CAP theorem in distributed systems, emphasizing how CockroachDB, a CP system prioritizing consistency over availability, navigates these constraints to achieve high availability. It argues that while the CAP theorem demands a trade-off between consistency and availability during network partitions, real-world scenarios often allow for systems like CockroachDB to remain highly available through strategic engineering practices. By maintaining strong consistency, CockroachDB ensures reliable data handling and minimizes complexity in application code, even though it might incur higher latency in some cases. The article highlights the introduction of bounded staleness reads in CockroachDB, which offers a balance between consistency and availability by allowing minimally stale reads during network partitions. Overall, the strong consistency foundation of CockroachDB provides developers with a reliable and straightforward data management experience, making it a preferred choice for many applications despite the potential trade-offs in speed.