Performance monitoring supports capturing numerical data from three different types of sources:
Windows performance counters and output from executables are monitored by the EventSentry agent running on the monitored machine, whereas SNMP counters are monitored by the Heartbeat Agent.
|
Trend and Leak detection are only available when monitoring Windows performance counters. |
Frequency Interval (Collect data every ...)
The frequency interval determines how often the performance counter values will be obtained/refreshed from the OS. Use low frequencies (< 5 seconds) for volatile performance counters (such as "Processor(*)\% Processor Time", "PhysicalDisk(*)\Avg. Disk Queue Length", ...), or when accurate data is needed. Use larger values for performance counters which change slowly (e.g. "Memory\Available MBytes"). Performance data collection is very efficient, and changing the interval will have little impact on the CPU utilization of the EventSentry agent. Still, it is considered good practice to select intervals based on the performance counter.
For SNMP counters, this value can not be smaller than the heartbeat polling interval.
Name
A descriptive name of the counter, this name will be visible in alerts and the web reports.
Treat data as floating point values
By default, performance counter values are interpreted as integer values, which is usually the best choice. Activate this option to force the performance counter values to be interpreted as floating point numbers (e.g. for performance counter values returning values smaller than 1).
|
When both a Windows performance counter and a SNMP counter are available, it is recommended to configure both in the same dialog. |
