I Work at Kohls.com and have noticed the following.
Note01: Testing Kohls.com production website
Note02: Clear Cookies & Cache for the past one day.
Webpagetest always gives number of requests to be around 180.
Chrome and Mozilla browser gives the number of requests to be around 240-300
We have two questions
a) Why is webpagetest requests so low consistently. (we are assuming that underneath you are using the same browsers)
b) Why is the number of requests constantly changing for the same website for the same day.
Do you have a link to a test result? Sometimes Chrome and Firefox will show resources served from cache as well as Data URI’s as requests but WPT only includes traffic that actually hits the network.
Presumably the number of requests change because there is some content on the page that is dynamic (in the case of kohls.com’s main page it looks like there is an adsense ad near the bottom for example which will be different every time).
pmeenan) Sometimes Chrome and Firefox will show resources served from cache as well as Data URI’s Sunil) We are hitting kohls.com using mozilla or chrome browser after clearing the Cache & Cookies which can be accessed from History menu. sunil) Our assumption is that underneath webpagetest, you are using Mozilla and chrome, so why are our requests consistently higher than webpagetest. pmeenan) in the case of kohls.com’s main page it looks like there is an adsense ad near the bottom for example which will be different every time sunil) This makes sesne, it still does not explain why sometimes we see a difference of 50 or more requests when compared to webpagetest, Does webpagetest not use adsense on kohls.com make to make requests. pmeenan) Do you have a link to a test result sunil) We can send you a Har file, is that acceptable.
Right, but depending on where you are looking for the information on the requests, it might be including things that do not actually hit the network (data URIs in particular).
Have you looked through to see what actual requests you are seeing that are not showing up on WebPagetest?
Sure, I was just explaining how the numbers can vary from one run to another and why they will not always be the same.
HAR file from the desktop browser or WebPagetest? Both need to be compared to each other to see what the requests are that are different.
Right, but depending on where you are looking for the information on the requests, it might be including things that do not actually hit the network (data URIs in particular).
Have you looked through to see what actual requests you are seeing that are not showing up on WebPagetest?
Sure, I was just explaining how the numbers can vary from one run to another and why they will not always be the same.
HAR file from the desktop browser or WebPagetest? Both need to be compared to each other to see what the requests are that are different.
[/quote]
Sunil: Comparing Har files which are turned into Excel sheets is in progress.
but here is a quick observation.
Webpagetests tells that there were 162Req (onLoad) & 192Req (pageLoad)
When we download the Excel sheet via the “RawObjectData” Link the number of lines is 265. We are assuming that these lines correspond to the number of requests. Can you please look into this. An attachment is provided for the Excel and a screenshot of the test.
The Raw Object Data includes both first and repeat view runs. There was a bug where I wasn’t including a column to indicate if the requests were from the cached run or not so I just fixed that.
Sunil:
Hi Patrick We have checked the Excel download from webpagetest. You have added an extra column called Cached with 0 & 1. We think it would be less confusing to have column name as RUN and its values as First or Repeat.