NetProfiler Analytics establishes a baseline of normal performance by tracking metrics over specific, recurring time intervals to understand typical behavior. It calculates this "normal range" by analyzing historical data for each specific time of day and day of the week. This methodology allows the system to accurately distinguish between a genuine performance anomaly and a predictable, cyclical fluctuation in traffic, such as the difference in network usage between a Monday morning and a Saturday night. The baseline is typically built from the previous four weeks of data for the same time slot and day.

A. Week of the month and month of the quarter: These time frames are too broad for the granular, short-term baselining used for real-time anomaly detection.

B. Month of the quarter and quarter of the year: This level of granularity is more appropriate for long-term capacity planning and trend analysis, not for defining normal operational performance.

C. Day of the week and week of the month: While "day of the week" is a correct component, "week of the month" is not the primary time frame used; "time of day" is the other critical factor for accurate baselining.

1. Riverbed SteelCentral NetProfiler and NetExpress User's Guide, Version 10.9.2, Chapter 10: Analytics, Section: "Performance Baselines", Page 229. The document states, "NetProfiler Analytics tracks the normal range for each performance metric for each time of day and day of the week. The normal range is based on the previous four weeks of data for the same time of day and day of the week."