I am not sure how the Cloudflare dashboard does it’s tests, but it might test after everything is cached on the Cloudflare edge CDN serve it is testing from. Also, this speed will depend on if you are caching html pages or not.
Remember, Cloudflare has 100s of CDN servers which may (or may not) serve your resources to clients, depending on if they are cached or not. In real life, it is likely that you might get some requests served from Cloudflare CDN (very fast, because their servers are likely very close to the visitor’s request), and some resources served from your origin server (which will be no faster than without Cloudflare).
These tests are arbitrary, and will depend on the visitors connection, the visitors location, and what resources are cached on the Cloudflare CDN for that specific visitors location.
Some of the best known 3rd party tools for measuring web speed and optimization:
However, even these tools are far from perfect. The won’t measure your website speed compared between with and without Cloudflare, and they will also focus mostly on the optimization of the output, which is only partially related to Cloudflare. They don’t necessarily cater as to how fast an actual human visitor will perceive your website.
In reality, you might not get the results you see in Cloudflare. It should be slightly faster, not slower, but this also depends on your physical location vs server location vs Cloudflare server location. You would need to test from a “private” browser window, with and without Cloudflare enabled.
Thank you! All you told is true, but it’s not the point here.
Cloudflare says to me that first paint time is 1 second.
BUT, all another web test tell me that this time is 4s or more. This time is exactly that cloudflare shows to me in my image above for “without cloudflare”
So, there is some aditional configuration to get cloudflate true velocity?
My actual plan is PRO.
*There is a way to configure cloudflare tests at São Paulo server? Now the test runs in california server
In my opinion, why rely on the Cloudflare report in the first place? I wouldn’t use that tool to objectively test my website speed with Cloudflare. Other tools are better and probably more accurate. Cloudflare’s test might be from a computer on Cloudflare’s network, and might have extremely fast network access to their own CDN servers. Also, their tests might run with everything that is cacheable cached, to mimic how fast your website might perform. In the real world through, it’s unlikely all your resources will get cached on CDN, and it’s unlikely other tools will load all resources cached.
I’m not quite sure what you are expecting in terms of “Cloudflare true velocity”. Your website’s main html document will always be served from your own server, and Cloudflare can’t speed up the delivery of this critical document. Cloudflare might speed up delivery of static assets like CSS, JS and images, and that might improve your page speed slightly, but then again that depends on those items actually being cached.
I use PRO myself, and there are some nice features, but to be fair, there is hardly anything in PRO that will speed up your website “paint time” more than the FREE version.
You mean the internal tests on their dashboard? I don’t think so. Personally I wouldn’t read to much into Cloudflare’s own report. It doesn’t have any info, and it’s hard to know how the tests are made.
that’s because the connection between you and your VPN is still a factor. Unless you physically move yourself/pc to a different geographic region, then your paint times are subject to geographic distance.
that is relative time not absolute and as @imagevuex mentioned depends on many factors
Including factors of the specs and network speed of the other pagespeed test site matter too.
Also Cloudflare Speed tab report has measure details link, click it and you’re still the geographical location CF is testing from. That should help too with relative tests.
example for me is measurement details link shows
Chrome | South Carolina, USA | 07/16/19 6:14:33
Connection Type: Cable
if your origin server happens to be hosted in Carolina USA, then chances of a faster page speed reported metrics behind Cloudflare is very high
FYI, Cloudflare Speed tab seems to be based on webpagetest too - maybe Cloudflare rolled their own private webpagetest instance as you can host your own private webpagetest instances
Example from webpagetest for my wordpress blog behind cloudflare hosted in San Jose, CA so US West Coast
I tested webpagetest at dulles US East vs dallas US Central vs california US West and as you can see first paint and render start is fastest with WPT tested from california US West as that is closest to my origin server location