Before raw data is purged, the data is summarized or “rolled-up” at predefined intervals. Roll-up data is the term given to the data that has been summarized. Before any raw data is purged, the minimum, maximum, and average values are calculated. These values can be preserved in several ways.
Hourly data is data that is summarized or rolled-up every hour. The raw data for each hour is processed to create a summary record for that hour. For example, if raw data is logged every five minutes, a total of twelve raw records are available in one hour. When data is rolled-up after one hour, the twelve raw records are processed to find the minimum, maximum, and average values during that hour. One summary record is generated for the twelve raw records. The hourly data is first computed on the agent machine. The data is then brought over to the server machine, where the data is added to the database.
The daily, weekly, and monthly values are calculated in a similar manner. The maximum, minimum, and average values are also calculated for each period.
Hourly data is kept for 7 days.
Daily data is kept for 5 weeks.
Weekly data is kept for 12 months.
Monthly data is kept for 24 months.