Answer by Florent D. ·
Hi Ian,
using our UEM terminology, a visit is a group of user actions for a single user.
A user action could be for instance one of the following:
- loading a page
- clicking on a button
Basically anything that requires a human intervention (use the mouse, keyboard).
Each action could comprise of several web request (html page+images for instance).
Visits are tracked separately for every single user.
I hope this helps.
Flo
Answer by Florent D. ·
my previous answer applies to UEM visits only (i.e via a cookie). Without UEM, you have to build measures to track SESSIONIDs so something similar.
Answer by Ian M. ·
Sorry I didn't see your full response so visits are identifed by cookies rather than IP. This is good to hear as I would expect a load test to be handled in the same way as real users in terms of unique visits. What is the visit timeout value and is this configuable?
yes you can configure it under your profile/user experience/pick the relevant application tab if you have more than one/"visit ends after user has been inactive for". The maximum is 60 minutes.
Best practice dictates that it should match your application session timeout.
Answer by Florent D. ·
Visits are unique for each user because we rely on a cookie and not the IP address.
The only case where you would see 2 visits for the same user would be if the user was inactive for a period of time (visit time out) and comes back later on.
Answer by Ian M. ·
Following on from my reply in performance testing terms a given load may be generated by only a few IP's (one per load injector platform) conmpared to hundreds of real users. Would this have any effect on how dynaTrace determines Visits and User Actions?
Most performance testing tools will not emulate a browser so UEM data will not be available (they just fire web requests without actually rendering them).
To get UEM data the injector must run javascript.
Answer by Ian M. ·
Hi Florent, Thanks for the information. I am attempting to use dynaTrace to come up with a load model for performance testing. I can tell the highest number of web requests in a given timeframe but I need to also understand how many unique visits and user actions resulted in the peak number of web requests. Is it safe to assume that each visit is orginating from a unique client IP or could the same IP be responsible for multiple visits in a given timeframe?
JANUARY 15, 3:00 PM GMT / 10:00 AM ET