4.55 SmartMemoryStats

Use this Knowledge Script to monitor the use of physical memory and swap usage of the system. When the usage crosses the threshold, the Knowledge Script raises an event indicating the memory usage with a detailed report of the memory utilization of the top memory intensive processes. The event also details paging scan rate information, KBytes swapped-in and swapped-out, and the page-out rate to determine if physical memory is a bottleneck. By providing all this information within a single event, this Knowledge Script eliminates the need to run multiple memory-related Knowledge Scripts and correlate the events from all of them to figure out what is going wrong, thereby overcoming problems like unnecessary filling up of database space, event storm, and event correlation.

A UNIX system with ZFS almost always displays all memory as used. The ZFS makes use of all the memory until another process needs it. Therefore, the Knowledge Script excludes ZFS memory usage for physical memory calculation and displays the memory that the processes actually use.

4.55.1 Resource Objects

Memory folder

4.55.2 Default Schedules

The default interval for this scripts is Every 5 minutes.

4.55.3 Setting Parameter Values

Set the following parameters as needed:

Description

How to Set It

General Settings

Event severity when job fails

Set the event severity level, from 1 to 40, to indicate the importance of an event reported when a job fails. The default is 5.

Enable debugging?

Select Yes to enable debugging. The default is unselected.

Event detail format

Select the format in which to view the event detail. The default is HTML Table.

Number of top memory intensive processes to display (0 for all)

Set the number of top memory intensive processes to display. To display all top memory intensive processes is set to 0. The default is 5.

Number of seconds between samples

Set the number of seconds, 2-30, between samples. The default is 5.

Number of times “vmstat” command should iterate before reporting an average value.

Set the number of times,1-100, that the “vmstat” command should iterate before reporting an average value. The default is 3.

Raise event when AppManager fails to get metrics?

Select Yes to raise an even when AppManager fails to get the metrics. The default is Yes.

Event severity

Set the event severity level, from 1 to 40, to indicate the importance of an event reported when a job fails. The default is 5.

HP-UX specific settings

Include reserved value in calculations?

Select Yes to include a reserved value in the HP_UX calculations.The default is Yes.

Include memory pseudo-swap values in calculations?

Select Yes to include the memory pseudo-swap values in the HP_UX settings. The default is unselected.

Event Settings

Raise event if physical memory exceeds threshold?

Select Yes to raise an event if physical memory crosses the threshold you specify for maximum percentage used. The default is selected.

NOTE:It is normal for UNIX systems to use almost all physical memory.

Threshold - Maximum physical memory used

Specify the maximum percentage (%) of physical memory that can be in use before an event is raised. The default is 90%.

Use swap usage in determining memory bottleneck?

Select Yes if you want to use swap usage in conjunction with physical memory usage before raising an event. The default is selected.

NOTE:If this option is selected, an event is raised only if physical and swap usage exceeds their respective specified thresholds.

Threshold

Specify the maximum percentage (%) of swap space that should be in use. The default is 90%.

Severity

Specify the severity level, from 1 to 40, to indicate the importance of the event when the maximum swap space usage crosses the threshold. The default is 5.

Raise event if swap usage exceeds threshold

Select Yes to raise an event if the paging file use crosses the threshold you specify for maximum percentage used. The default is unselected.

Threshold

Specify the maximum percentage (%) of swap space that should be in use. The default is 90%.

Severity

Specify the severity level, from 1 to 40, to indicate the importance of the event when the maximum swap space usage crosses the threshold. The default is 5.

Collect data settings

Collect data for physical memory used?

Select Yes to collect charts and graphs for physical memory use. The default is unselected.

Collect data for percentage of computation memory in use (AIX only)?

Select Yes to collect charts and graphs for the percentage of computation memory in use for AIX systems. The default is unselected.

Collect data for total virtual memory used?

Select Yes to collect charts and graphs for the total virtual memory used. The default is unselected.

Collect data for swap space used?

Select Yes to collect charts and graphs for the total swap space used. The default is unselected.