Investigate these checkout interaction failures using the approved, redacted browser events, traces, support report, and release diff. Compare affected and unaffected journeys. Propose competing explanations, then reproduce the most plausible one in staging with synthetic data. Don't interact with customers or change production. Show how the result should change our regression coverage and monitoring.