Microsoft Exchange 2000 Operations Guide — Version 1.054
Performance Monitoring
Performance Monitoring is the monitoring of existing system(s) to ensure that optimum
use is made of the hardware resources, and that agreed performance levels can be main-
tained.
Performance Monitoring allows you to determine if your server running Exchange 2000 is
meeting the performance standards you have defined in your service level agreements
(SLAs). Over time, you can use Performance Monitoring to generate data that can be used
in trend analysis. This alerts you to possible performance and availability issues in the
future, and allows you to solve problems before they arise).
One of the first tasks involved in performance monitoring is to generate a baseline. This
baseline is a measure of what figures you expect to see when measuring a healthy system.
This can then be compared to the figures you gather in day-to-day monitoring, allowing
you to track problems easily.
In this section, you will look at the objects and counters that you may want to monitor
using System Monitor. These parameters will form the basis of your baseline. Y ou will also
examine centralized monitoring techniques for remote servers.

System Monitor

If your e-mail system was Exchange Server 5.5, you are probably accustomed to using
Microsoft Windows NT 4.0 Performance Monitor to analyze the performance of your
Exchange 5.5 server. Exchange Server 5.5 includes a series of Performance Monitor
Workspaces to allow you to quickly see in graph form a series of key counters.
The Windows 2000 operating system includes System Monitor (which consists of Perfor-
mance Monitor and Network Monitor) for analyzing the performance of your system.
When you install Exchange 2000 Server, a large number of objects are installed and
counters are associated with those objects.
It is worth noting that while real-time graphs created in System Monitor often look
very pretty, they are only of limited use, particularly if no one is looking at them. If you
continually monitor 500 different counters on your server running Exchange, the self-
monitoring uses CPU cycles. You have now undermined the performance of that server just
by monitoring it. So only monitor what you need to, and consider using Performance
Logging and Alerts, which can produce much more useful information with less of a load
on the server. Reducing the frequency of monitoring produces much less of a load on the
server and in many cases produces a more accurate picture, depending on the counters in
question.
Note: Remote monitoring is almost always better than self-monitoring, because performance
is not tainted by the load caused by monitoring. For more information about remote monitoring,
see articles Q243283, Creating a Log File to Send to Customers for Remote Monitoring and
Q240389, Error Message: Event ID: 2028 “The Service was Unable to Add the Counter
\\Server_Name\Counter_Name” in the Microsoft Knowledge Base.