From: Chip W. <chi...@hy...> - 2007-03-09 21:52:25
|
Much of the information you are looking for can be found in our online documentation here in the Data Manager Configuration Properties section: http://support.hyperic.com/confluence/display/DOCSHQ30/HQ+Server+Configuration The short answer is, HQ stores monitoring data using a tiered model to reduce the amount of data that is stored and still provide the necessary granularity to determine when monitoring data changes. By default HQ will keep detailed metric data as defined by the metric collection intervals for a period of 2 days. After 2 days the data will be compressed and archived into hourly averages with highs and lows. This is very similar to data storage techniques used by other tools you might already be familiar with, such as RRDTool. The net result is slow, controlled growth of the database size while still allowing you to get good performance data spanning a considerable period of time (including your % uptime for the year example). You can modify how long data is held, and how often clean-up occurs for various types of data through Administration -> HQ Server Configuration. |