The ebb and flow of "web performance"

Applying eValid to Rich Internet Application (RIA) Performance Monitoring.

The ebb and flow of "web performance"

Postby oMyles » Wed Feb 20, 2013 1:16 pm

How much influence does the ebb and flow of "web performance" have on end-user performance measurements?
Posts: 1
Joined: Wed Feb 20, 2013 1:14 pm

Re: The ebb and flow of "web performance"

Postby eValid » Wed Feb 20, 2013 11:25 pm

If you mean by "ebb and flow" the natural tendence of web response times to vary in a regular way with the time of day, you're right. The delivered end-user performance will go up and go down in the same proportion.

What we've observed, interestingly enough, is that the intra-event timings in a typical Rich Internet Application (RIA) test playback tend to remain relatively constant. It seems as if the "startup" time is what varies the most. Once the connection is made (assuming the waits in the script are not so long as to exceed the "keep alive" time of the HTTP connection) then the only thing that variances internal to the application shows is variation due to response time at the server.

That's why we make such a big thing out of having the abiity to threshold inter-RIA-test timings: this is the way to separate the normal ebb and flow of web performance from the abnormal behavior that may be exhibited by your server.

Of course, THAT behavior is where the payoff is in RIA monitoring.

eValid Monitoring Team
Posts: 1853
Joined: Tue Jan 01, 2008 12:48 pm
Location: USA

Return to Rich Internet Application Monitoring (RIA)

Design Downloaded from free phpBB templates | free website templates | Free Web Buttons