The basis of all reporting and dashboard presentations is retained data from established monitors. In other words, each monitor provides a simple schema from which you can produce a chart, graph or report.
To reduce resource impacts, the scope of retained data may exclude some of the collected data. A monitor may have no retained data and only emit events based on transient results in the execution/calculation.
For example, the application can derive a metric from several collected values and you may opt to retain only the derived result.
All monitors rely on a polling engine which provides runtime mediation activities for distributed device interaction at regular intervals. Monitors may share a retention policy. Data is rolled up hourly and daily into summary data. The retention policy controls how long data is held per roll-up period. You must select the correct period see what has been collected.
When you manage these policies, you configure how monitored data is retained. When you select Manage Retention Policies in the Monitors portlet, first a list of available policies appears.
Clicking the Add button at the top of the screen lets you create a new policy, while clicking the Edit button to the right of selected, listed policies lets you modify existing policies. The Delete button to the right of listed policies removes them from the list.
Monitors may share a retention policy. The retention policy controls how long data is held per roll-up period. The editor for Retention policies lets you assign characteristics and monitors to them.
The editor contains the following fields:
General Retention Policy Options
Policy Name -- A text identifier for the policy.
Description -- An optional description for the policy.
Detail / Hourly / Daily Data (Days) -- How many days to retain the selected data.
The amount retained has both a performance and data storage impact. For example, retaining day’s information from an active performance SNMP monitor configured with one target’s worth of data, retrieved on one minute intervals can consume 0.7 G of database, and require 21 inserts per second.
Traffic flow analysis can process and retain even larger amounts of information. Flows that correlate 50%, polled every minute for a day require roughly 109G of database, and require 4500 inserts per second.
Active Monitor Members
Select from Available Monitors on the left, and click arrows to move the desired monitor(s) to the Selected Monitors on the right.
Click Save to preserve your edits, and include the monitor as listed among existing Retention Policies, or click Cancel to abandon any changes.