> From: Tony Fountain [mailto:[EMAIL PROTECTED]
> Would you happen to know of any good references that walks through how
> to trouble shoot Apache related performance issues?
I don't, so I'm going to throw that open to the list here :-).
A few obvious pointers from me:
- Performance Monitor: %CPU, Memory pages/second, average disk queue
length. Over 80% sustained, over 100, or over 2 per spindle in your
disk array are "finger in the air" figures for bottlenecks on that kind
of resource.
- Remember to monitor your database server if it's separate - and the
traffic across your net link to/from the RDBMS. I've seen plenty of
people scratching their heads over an almost-idle Web server when the
database server was thrashing away.
- Turn on verbose GCs in the Java VM and watch for significant GC
activity. This can also show as spikes of CPU usage, and generally
indicates you're short of Java heap space.
- The single most common cause of poor webapp performance, in my
experience, is poor use of the database. It's easy to send thousands of
small queries to the RDBMS - at which point you'll be crippled by time
to parse and execute all these queries, plus latency if the RDBMS is on
another machine. It's also easy to create big queries that make poor
use of indexes, and saturate the RDBMS that way. In general, more
hardware doesn't solve the problem; sending the person who built the
database and/or webapp on a course on RDB optimisation is the way to go
;-).
- Peter
---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
