HP Insight Control Performance Management 7.3 User Guide

The Analysis Explanation details an above-average page read rate (higher than 30% of the hard
page fault), and less than configured memory threshold. The analysis recommends more physical
memory to handle the load. Such recommendations may suggest adding hardware.
In this example, the situation might have occurred because a new batch job was assigned to run
on the server at night.
Displaying the memory graph
To show a graphical display of memory counters in the Online Analysis pane, click the Graph tab.
The Graph pane appears. See the following figure:
This graph displays the number of samples as selected in Monitoring Administration page. The
duration of the graph would be for a period that is the same as sample rate (default 120) multiplied
by Number of Samples.
The pattern of the graph shows rising, peaking, and declining values, which might be caused by
the completion of the new batch job. At the time the graph is created, the batch job no longer
causes a problem. However, the performance information is logged in the database for later review
with the Offline Analysis tool.
Analyzing a virtual machine host and guest bottleneck condition
The following sections discuss the actions to take when a bottleneck exists on virtual machine hosts
and guests.
IMPORTANT: The virtual machine host must be licensed for HP Insight Control. It must be registered
on the CMS before being licensed and monitored by performance management.
Selecting the server (virtual machine host)
Click the PF column against the node you want to analyze to display the performance management
Online Analysis window.
The screen displays the server node tree in the left configuration navigation pane and the Status
tab in the right pane. The counters that appear are selected items from the various components.
See the following figure:
70 Bottleneck scenarios