Thresholds define what is "normal", "warning", and "critical" for each KPI. ITSI supports static thresholds (fixed values) and adaptive thresholds (dynamically calculated based on historical patterns). Adaptive thresholds use machine learning to detect when a KPI deviates from its expected behavior.
Time Policies allow you to define different thresholds for different time windows. For example, CPU utilization of 90% at 3 AM during a batch job might be normal, but the same value at 2 PM during business hours is critical. Time policies prevent false alerts caused by predictable workload patterns.