Time to First Byte: F!

We are preparing to deploy a nopCommerce site (http://www.trishoponline.com) and I've followed many of the optimization guides meticulously, deleting unused Languages and Countries and deactivating plugins to improve site performance. We've got a few images left to optimize, but otherwise I'm pretty comfortable with its performance, except in one area.

Using webpagetest.org, my "Time to First Byte" is atrocious, at 1+ seconds at its best and sometimes more than 2 seconds. I'm using a cloud server with 2 GB of memory (and plenty is still available) and a dedicated SQL Server, both on Everleap. I'm considering a CDN in the future but ultimately, performance once the site responds and starts downloading the page and images, things are satisfactory.

Test yourself if you like:

https://www.trishoponline.com/2018-cervelo-p3-8060

Anyone else experienced similar issues and corrected them? Any guidance on dealing with performance issues like this? This just seems like low hanging fruit if I can get the TTFB down under a half-second.

TTFB is a challenging issue - at a big half billion dollar company I managed web development for we had a hard time getting our TTFB down below 2 seconds on an MVC 5 application.

I do feel that one significant performance improvement would be to get away from EF for data access - utilize stored procs and optimized queries more and get rid of the overhead of EF. Problem with this route is once you replace the data layer you are stuck in a challenging spot for upgrade path.

Just a little trick..

Open shared/_productbox.cshtml..

Remove all widget call you find in it (sure, if you dont use it).

See, if you have 30 products on a category page, for each product box nop will check at least 4 times for a widget.. even in cache, there are 30 x 4 x XXms, and it is a waste of time, literally.

Another thing to check.. avoid using too many plugins, nop cant handle to much plugins very well and the result is bad TTFB too.

iob2000 wrote:


Thank you for that tip!!!

I just tested it.  

A page that uses the shared/_productbox.cshtml had a 2.08 TTFB and a Lighthouse speed score of 31! Ouch!

After removing all the HTML.Widget instances, TTFB was fixed and the Lighthouse speed score jumped to 50.

I will now go though all the razor views and remove the unused HTML.Widget instances.

Thank you!