I’m like to demonstrate piwik to my boss in a way more close to our reality. For that I’d like to simulate a few hundred users in a real case.
Does anyone ever did something like that?
For that I’m using jMeter. Since jMeter works on the HTTP calls only (it does not execute javascript content) I “recorded” a live web session with a proxy and I’m trying to just change the parameters passed over to the piwik.php page.
My first problem is: I’m having problem to made calls as different users. How and when does piwik set the cookies that it uses to identify a single user?
There is other questions, but I’ll ask them in different threads.
Is the OS, Browser, Screen res, or anything is changing while your script is running ?
Piwik use gathered data about user config and all and hash it to make a unique identifier.
Maybe it also place a cookie, that i dont know.
All i know is that if i test a page in different browser, each test is considered a different unique user (and a returning user after the first hit).
[quote=TulipVorlax @ Feb 22 2010, 11:41 PM]Is the OS, Browser, Screen res, or anything is changing while your script is running ?
Piwik use gathered data about user config and all and hash it to make a unique identifier.
Maybe it also place a cookie, that i dont know.
All i know is that if i test a page in different browser, each test is considered a different unique user (and a returning user after the first hit).[/quote]
No. Below is the data that my tests are passing. Data enclosed in curly braces are dynamic (jmeter changes it before sending), the rest are “constant” values:
idsite=2 url=www.mysite.com
res=1024x768
h=<random number between 0-23>
m=<random minute between 1-59>
s=<random second between 1-59>
cookie=1
urlref=
rand=0.029115581872977803
pdf=1
qt=0
realp=1
wma=0
dir=0
fla=1
java=1
gears=0
ag=1
action_name=