Clock synchronization is a critical challenge in distributed systems like CockroachDB, which use the Network Time Protocol (NTP) along with software-based compensation techniques to achieve reliable time coordination across database nodes. Good timekeeping is essential for maintaining ACID guarantees and preventing stale reads. Clock skew between cluster VMs can impact SQL workload response times, generally degrading query performance as larger skews occur. CockroachDB relies on the clocks of all its VMs being within a configurable maximum offset from each other to provide non-stale reads. The max-offset value determines the uncertainty interval for reading transactions and influences the probability of contention, with smaller values leading to more retries. Time smoothing of leap seconds is crucial, particularly in regions where NTP sources implement smearing or slewing. Live VM migrations can support two operational tasks: runtime VM load optimization and memory-preserving maintenance, but require uninterrupted clock continuity to guarantee transactional consistency. CockroachDB operators must develop custom procedures for handling planned maintenance events, including disabling live migrations when uninterrupted clock is not available, and manually triggering VM maintenance after stopping nodes. Precise clock management is a foundational requirement for maintaining the correctness, performance, and reliability of CockroachDB in distributed environments. By following best practices outlined in this article, organizations can significantly reduce the risk of transaction retries, node instability, and stale reads, resulting in a more resilient and performant deployment.