WebPagetest Forums
Start Render 0.000s - Printable Version

+- WebPagetest Forums (https://www.webpagetest.org/forums)
+-- Forum: WebPagetest (/forumdisplay.php?fid=7)
+--- Forum: Bugs/Issues (/forumdisplay.php?fid=10)
+--- Thread: Start Render 0.000s (/showthread.php?tid=12190)



Start Render 0.000s - yshemesh - 04-17-2013 06:59 AM

It seems that whenever hitting a webpage that results in non HTTP/200.
(like 400 Not found page - which could have its own customized page with images, etc....)

The Start Render time reported is always zero.
This was tested with the public system hosted at webpagetest.org


RE: Start Render 0.000s - pmeenan - 04-18-2013 11:41 PM

It's not only the start render that is 0, all of the times are invalid when you get a page-level error back. WebPagetest doesn't load custom error pages and fails-out as soon as the browser indicates that the page failed.


RE: Start Render 0.000s - yshemesh - 04-19-2013 05:47 AM

(04-18-2013 11:41 PM)pmeenan Wrote:  It's not only the start render that is 0, all of the times are invalid when you get a page-level error back. WebPagetest doesn't load custom error pages and fails-out as soon as the browser indicates that the page failed.

Thank you Pat.
I know this is not a common use-case, but still many customized HTTP/404 web pages incorporate full page graphics and styling. Ability to benchmark those could perhaps be regarded as a valid use-case.

So if this is "by design" just as conditioned in file PHP "page_data.inc" (// only look at non-error runs) it would be easy to resolve, but when we done so I still see the "Start Render" time as zero (other times are Valid non Zero), so probably the agents are not getting the right metrics.
If that is the case and it is a technical limitation, I guess it will be harder to resolve.


RE: Start Render 0.000s - pmeenan - 04-19-2013 06:04 AM

It is done on the test agents themselves. Do you have a way to directly access your custom 404 page (like /404.html)? Otherwise, a blank document with an iframe pointing to a non-existent source would work as well.