Extra request to rapidssl

Support
1

Hello. I’ve noticed that whenever i run a test by firefox or firefox nightly to https protocol website, test result waterfall has a request to http://rapidssl-ocsp.geotrust.com/

As far as i know the target first time byte is calculated, based on times of first request, therefore giving wrong results.

here is an example:
http://www.webpagetest.org/result/141231_1T_XPJ/1/details/

webpagetest says that expected FTB is 87ms, although it should be 173ms(SSL)+3*76ms(connect time) = 401ms.
or
28ms(DNS)+173ms(SSL)+76ms(connect)+100ms=377ms.

But surely not 87ms.

This messes up a lot of results, and, from client point of view, shows that optimizations made load times worse (judging by grades). Which is not cool.

If i’m missing a part of brain - please, advise :slight_smile:

Thanks.

2

Sorry, you’re correct - that’s a bug. I opened an issue to track it: TTFB incorrect when OCSP/CRL checks are done · Issue #383 · WPO-Foundation/webpagetest · GitHub

That said, you should be using ocsp certificate stapling so the browser doesn’t have to make the external request to check the certificate.