> I was wondering what was the general pratice for testing
> your app is....

In my experience, the general practice is to release your application, and
it either supports the number of users it needs to, or it doesn't. Of
course, that's not the best practice, but it's certainly the most common.


> Is there any way of similating a certain number of hits
> per minute?

Yes, there are plenty of load testing tools. Some are free, such as my
favorite free load tester, OpenSTA (http://www.opensta.org/). Some are
incredibly expensive, such as Segue SilkPerformer, which can run up to $50K.
The commercial tools generally provide richer feature sets, and if you do
enough load testing, they're worth the money in the amount of time they save
you.

> I'd like to test some sections of some apps and see what
> kind of performance my code is getting (or not getting).
>  
> Does anyone have any suggestions as to how this test could
> be done?

Unfortunately, load testing isn't a trivial task. In general, though, to
oversimplify things a bit, you write a test script which travels through
your application, then you give that script to your load test tool and let
it run the script with some predetermined number of virtual users. As it
runs the script, it measures response time and throughput, and stores those
measurements. After the test has been run, you can generate data about which
pages took longest on average, and those are your bottlenecks.

You can then examine those pages specifically, and use GetTickCount within
those pages to find out which section of the page is causing trouble.

This kind of load test is called a performance test. You can also use load
test tools to perform capacity planning tests.

> I liked the example on the houseoffusion.com home page where
> the question was asked, "Is CFLOOP really needed?"... little
> tricks like that, I'm sure are almost trivial but running
> sloppy queries/loops....etc... Just want to tighten it up.

I'm sorry, but that kind of question is like asking how many angels can
dance on the head of a pin. You won't find your performance problems can be
solved by "little tricks" like that. I've analyzed literally hundreds of CF
applications, and their problems were never caused by those sorts of things.
Instead, the problems almost always come down to poor database schema
design, poorly written SQL, inadequate caching, or some combination of
those.

Dave Watts, CTO, Fig Leaf Software
http://www.figleaf.com/
phone: 202-797-5496
fax: 202-797-5444
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings] [Donations and Support]

Reply via email to