[ILUG] Slightly OT: Suggestions for performance testing PHP website

Phil Bradley philb at vodafone.ie
Mon Dec 15 16:36:26 GMT 2008


Hi all,

I just noticed a few replies had come in after I thought the thread had 
finished, sorry if it looked like I was ignoring these posts.

Firstly, for everyone who pitched in with suggestions, thank you very 
much. <plug>The site is now live and is at http://www.icty.org. I can't 
claim this was all my own work but I provided the development services 
required</plug>.

It's still a little sluggish at approx 25 requests/second (even with 
XCache enabled) but good enough for the moment. I'll have a look at 
wincachegrind, as per David's suggestion. Since none of the included 
files actually *do* anything apart from define classes and functions, it 
looks like this is just a case of the server just not being very fast 
(it's running in a VM which might be under resourced).

Philip: I'll probably revisit JMeter, it may have just been a mental 
block but I wasn't confident that I understood what it was reporting. 
The ApacheBench report seemed to make more sense and, while less capable 
than JMeter, it sufficed for a simple comparative benchmark. Thanks for 
the suggestion though.

Cheers all,
  -Phil






PhilipQuinlan wrote:
> Hi Phil,
> 
> not wanting to start a row over our favourite tools, but  in what way 
> did JMeter fail to do what you want, it should be perfect for what you 
> describe.
> 
> Build individual application function tests, string them together 
> randomly, top and tail with login and logout then run n threads.
> 
> It sounds ideal for your requirements.
> 
> Philip
> 
> 
> 
> __________ Information from ESET NOD32 Antivirus, version of virus 
> signature database 3660 (20081203) __________
> 
> The message was checked by ESET NOD32 Antivirus.
> 
> http://www.eset.com
> 
> 
> 
> 




More information about the ILUG mailing list