Navigation:  Monitoring with EventSentry > System Health Monitoring > Performance Monitoring >

Database Tracking

Collecting performance data in a database allows you view the current performance status and performance data history through the web reports, using either graphical charts or output in HTML and CSV format. Database tracking is also required when using the dashboard feature of the web reports.

EventSentry gives you flexibility by allowing you to configure which counters are written how frequently to the database.

To record performance data in a database you can either:

Adding database tracking to an existing counter

To add database tracking to an existing counter, simply double-click the counter in the performance tracking dialog and check the "Database" checkbox. See below for more information.

Creating a new counter for database tracking

Follow the instructions in Alerts to create a new counter, you can ignore the section "Enable Alert" if you do not wish to receive alerts on the counter and only log to the database.

The name, counter and polling interval are required when recording performance data in a database. Please see alerts for more information.

Log to ODBC action

Check this checkbox to record performance data in a database and select an ODBC action.

Logging Interval

The logging interval defines how often the performance data is written to the database. The logging interval has to be equal or higher than the polling interval, since the polling interval defines how often the data is actually being retrieved from the OS.

Log Average

"Log Average" is only useful when the polling interval is smaller than the database logging interval. Activating this feature will instruct EventSentry to write the average measurement of all polling data to the database, rather than writing only the value that was measured at the time data needed to be written to the database.

This feature is best explained with an example. Let's say that you are recording the average CPU utilization to a database. The polling interval is 10 seconds, and you configured the logging interval to 1 minute. Every dot in the graph below represents data collected at the polling interval.

If "Log Average" is not checked, then EventSentry will write the current measurement of 46% to the database at time interval 6, and the current measurement of 1% to the database at interval 12. As you can see from the graph above, neither one of those two values accurately describes the actual data.

If "Log Average" is checked, then EventSentry will calculate the average from 6 polling value, and log that information to the database. In the example above the first value to be written will be 19% (instead of 45%), and the second value will be 9% (instead of 1%).

We recommend using a polling interval of around 10 seconds and a database logging interval that suits your needs, for example 10 minutes. Keep in mind that a small DB logging interval will give you very accurate reports, but will use up significant amounts of disk space, especially in large networks. Using the "Log Average" feature allows you to write only a minimum amount of data in the database.