http://h21007.www2.hp.com/portal/download/files/unprot/devresource/Docs/TechPapers/UXPerfCookBook.pdf
http://www.ablongman.com/samplechapter/0130187062.pdf
Glance and Performance Agent (PA is also known as MeasureWare) have the concept of Blocked States (which are also known as wait reasons). You can select a process in Glance, and then get into the Wait States screen for it to see what percentage of time that it’s waiting
for different resources. Unfortunately, these don’t always point you directly to the source
of the problem. Some of them, such as Priority, are easier: if a process is blocked on
Priority that means that it was stuck waiting for CPU time as a higher-priority process
ran. Some other wait reasons, such as Streams (Streams subsystem I/O) are trickier. If a
process is spending most of its time blocked on Streams, then it may be waiting because a
network is bottlenecked, but (more likely) it is idle reading from a Stream waiting until
something writes to it. User login shells sit in Stream wait when waiting for terminal
input.
We’re focusing on performance, not performance metrics
If you have Glance on a system, run xglance (same as gpm) and click on the Help -> User’s Guide menu selection, then in the
help window click on the Performance Metrics section to see all the definitions.
Alternatively, in xglance use the Configure -> Choose Metrics selection from one of the
Report windows to see the list of all available metrics in that area, and you can right-click
to conjure up the metric definitions.
It is not always easy to find out why the CPU bottleneck is happening. Here’s where it is
important to have that baseline knowledge of what the system looks like when it’s
running normally, so you’ll have an easier time spotting the processes and applications
that are contributing to a problem. Stephen likes to call these the ‘offending’ process(es).
