Hi,
I’m looking at my results here: http://www.webpagetest.org/result/110509_BA_JK20/1/details/
I see that there is a space of no activity.
Can anyone tell me what this is?
Thanks,
Shamil
Try more runs or from a different location just to verify it is real. It doesn’t look like the first view test went very smoothly (shorter waterfall than the repeat view and no screen shot). I re-ran the test and got a result that looks more like I would expect: http://www.webpagetest.org/result/110509_FW_JK30/
It’s possible we still have some kinks to work out with the Chrome support (though this is the first time I have seen that kind of an issue).
Thanks,
-Pat
Hi Pat,
That does look much better.
Onto this one: http://www.webpagetest.org/result/110509_WX_JK5G/.
Ouch first page load isn’t great. Time to remove the extra JS.
Wow - awesome example of why to merge css and js files 
Have you tried W3 Total Cache? It can help you merge the files together (you need to tell it which files to combine but it will take care of the rest).
I’ve seen a similar issue regularly with the IE8 machine out of Virginia. Other IE8 machines are fine and VA IE9/Chrome are also fine.
http://www.webpagetest.org/result/110510_V4_JZ74/
So, these “no activity” reports can be normal. They may be a result of slow-executing javascript files which can hurt perceived performance (the images are loading in the background after the page has downloaded).
That is REALLY bizarre. I wonder if there is something triggering an IE8-ism. I wouldn’t think it was slow Javascript because the CPU utilization drops to 0 (it normally spikes to 100% for javascript issues).
We actually used to see the delay with no activity on our website from different machines about 10-25% of the time. We had a ShowPage script which would only show the page once all items are downloaded. Once we removed that script, then we saw the page load immediately (even while the coremetric tags were loading in the background). Now we cannot reproduce this 15 second ‘no activity’. But I still see the issue with WebPageTest IE8 out of VA. We flushed our CDN. Is it possible that the Virginia machine is still caching the cookies or site content?
No, the cache gets completely nuked between test runs (including the index files). Did you change file names or is there any way to detect the new/old code from just the requests?
Yes, since the 404 errors are gone, it does appear that the IE8 is hitting the correct code. Interestingly, I am impressed that WebPageTest correctly caught the period of no activity that I can see with FF4 & IE8. Keynote was unable to catch this period of inactivity. But now I have to wonder why I can no longer see it but this site is. I have tested with/without the CDN. Now I’m going to have to find a Virginia proxy server to see if one of our geo tags are off.
Let me know if there’s any tests you want me to run that can help figure it out. Worst case I can probably also set up RDP access on a test machine if you want to poke around.
FWIW, Keynote has multiple different metrics and the default is usually “network time” (at least it was when I used to use it) and that removes any gaps from the time. There should be a “user time” or something like that which is the wall-clock time for the test.