Setting thresholds for resource utilization that trigger alerts before reaching capacity ensures that the operations team is proactively informed about potential performance issues. By receiving alerts before the system reaches its full capacity, the team can take preemptive actions to mitigate any impact. This aligns with best practices for maintaining a high level of application performance and user experience. Setting the thresholds at the exact capacity or only logging the incidents would either result in a reactive approach or in no alerting mechanism to preemptively address the issue. Monitoring without alerting would not facilitate immediate action since the degradation could go unnoticed until reviewed.