Process variation is fluctuation in outcomes caused by changing conditions in the system. In software teams, a metric such as lead time is most useful as a signal for learning and improving the process—not as a score for ranking individuals. Blaming a person for a system-driven fluctuation can suppress the very signals that would help fix it. section 9.1.1
A cycle-time fluctuation
A team notices cycle time rising every other week. Investigation links the pattern to senior developers rotating onto on-call duty and having less time for code reviews. The team schedules review windows around those shifts, and cycle time stabilizes. The metric points to a process condition the team can change. section 9.1.1
In production, fluctuating demand or request-processing times can expose a lack of capacity margin. A service at high steady-state utilization may fail to absorb a traffic surge, causing response times to spike and timeouts to cascade. Headroom buffers variation; uneven demand makes that buffer more important. section 12.5.2
Figure 12.10 Headroom determines whether a traffic surge is absorbed or becomes a cascade.

Use the figure to trace how limited headroom can turn a traffic surge into cascading failures. section 12.5.2
