Perfmon? Haven’t we been here before? If you aren’t aware of the historical performance of your servers, then how can you compare today’s “it’s slow” with the “it’s slow” from yesteryear?
If you haven’t yet enabled performance counters on your servers, consider enabling at least a few critical ones so you can keep a baseline of the performance of your servers. Some of the critical ones to keep are:

  • Disk\% Free Space
  • Disk\% Disk Time
  • Disk\Current Disk Queue Length
  • Disk\Disk Reads and Writes/Sec
  • Memory\Available Mbytes
  • Memory\Pages/sec
  • Page File\% Utilization
  • Processor\% Processor Time
  • Processor\Interrupts/sec
  • System\Processor Queue Length

Logman can configure nearly every setting for a Perfmon counter set.
Use it with the following syntax:

Logman.exe VERB {collection name} {options}

The collection name is the name of the Perfmon counter set you’re
interested in creating or manipulating. For the VERB, choose from

Create [countertrace] — Create a new collection.
Start — Start an existing collection and set the begin time
to manual.
Stop — Stop an existing collection and set the end time to manual.
Delete — Delete an existing collection.
Query — Query collection properties. If no collection_name is given
all collections are listed.
Update — Update an existing collection properties.

Like with file shares on a server, there is a specific syntax to Perfmon
counter addressing. For a simple counter like % Processor Time on the
local system, the proper addressing is:

\Processor(_Total)\% Processor Time

For more complicated counters that include object instances and remote
machines, the addressing can look like this (it’s line-wrapped here,
but you’d type this as a single line):


An example of a counter that includes more of the elements above would
be to measure % Processor Time for a particular Terminal Services
session on a remote computer:

\\Server1\Terminal Services Session(RDP-Tcp 3039)\% Processor Time

Where I have found the best use of this tool is in starting and
stopping counters through scheduled tasks. Performance counters suffer
from averaging effects when they run at times no one is using the
server. To prevent nighttime lack-of-use from negatively skewing your
results, use logman with the “start” and “stop” switch from a scheduled
task to relegate your performance measurement to only your
business hours.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: