Question Wordpress+Woocommerce TTFB

[jojo]

Hi guys,

Im running a wordpress + woocommerce website on a VPS, and I'm struggling to get a decent TTFB out of it.

This is without any traffic load on it! 0 users.

I'm serving nginx, php 7.3 with FDM, GZIP compression.

I tried everything you can think of without avail.

With all plugins turned on (w3 caching, autoptimize, woocommerce, ...) I get an average TTFB of 5s!

When I disable and delete all plugins (including woocommerce) it improves a bit down to 1.5/2 seconds, but still a huge conversion killer.

I also tried with and without cloudflare, same performance.

Here is the VPS stats:
CPU Intel Xeon E312xx (Sandy Bridge) (2 core(s))
Version Plesk Obsidian v18.0.21_build1800191128.17 os_CentOS 7
OS CentOS Linux 7.7.1908 (Core)
2GB memory

Not sure at this point what might be causing this, VPS benchmark shows io speed of 75mb/s, so disk speed is also not that great on this VPS.

Are both CPU cores used here by default? not sure how plesk handle this, but is my wordpress installations utilizing both cores?

 

[IgorG]

Have you tried recommendations from TTFB for site in Plesk is too high. What can be done to improve it? ?

 

[jojo]

Yes thats the first thing I tried, I applied all the recommended settings from that page. Still slow TTFB. Not sure how I can identify the bottleneck.
Shopify websites have an average 0.2s TTFB, and that with a fully loaded website with several plugins, mine has a bare bone woocommerce installation and still about 5 times slower. Not sure if a traditional hosting environment can get close to that.

 

[Martin.H]

Shopify websites

Well you cant compare Shopify with WooCommerce,
Wordpress/WooCommerce is very bad when it comes to caching. The best way to cache there is statically, but you pretty sure do not want to cache a Shop statically. Thats what makes WooCommerce so bad.
It was never ment to be a shop, but a Blog CMS. I would recommend you to switch to Shopware (6) which is free and very very awesome! But to be honest a TTFB of 1.5/2s is very slow and I think I can remember that I had much better TTFBs even with WooCommerce.

 

[seqoi]

Well you cant compare Shopify with WooCommerce,
Wordpress/WooCommerce is very bad when it comes to caching. The best way to cache there is statically, but you pretty sure do not want to cache a Shop statically. Thats what makes WooCommerce so bad.
It was never ment to be a shop, but a Blog CMS. I would recommend you to switch to Shopware (6) which is free and very very awesome! But to be honest a TTFB of 1.5/2s is very slow and I think I can remember that I had much better TTFBs even with WooCommerce.

Your post is an absolute misconception and you are spreading wrong information. Without going too much into details (i need to some degree) i would advise in friendly manner that if you don't know what you are talking about you simply do not post such false info because

a: you can actually lead OP in very wrong direction.
b: someone later can read what you wrote and again damage can be done.

Before i continue allow me to tell shortly that we actually specialized big part of our business in Wordpress or Woocommerce. We where at Magento and Joomla years before that. So i do know what i am talking about. And i will provide you some real world values.

Yes Wordpress started as a Blog platform. But that was 16 years ago. Number Six and ten. You would logically think that something happened in this 16 years period causing Wordpress (and Woocommerce) to become one of the most popular CMS to this date. Flash news - it's not because of Blogs which are dead anyway. So what it could be?

Some real world values from me. Hetzner virtual server. CX21 (non dedicated cores) Truly thrifty cloud hosting - Hetzner Online GmbH - running Obsidan and 27 individual wordpress websites in various subscription models.

Client 1. Wordpress + Woocomerce. Wholesale clothing. more then 18 000 articles with images. Several gigabytes of data (16gb i believe) + realtime sync of items and their stock values with server in another country. TTFB varies between 100 to 400ms. GTmetrix gives 79% (but that's relative anyway), Pingdom gives 87%. - Same client was on supposedly superior Magento 1.9 before we migrated him. His whole business is actually on another level now. Client is so satisfied that we got different contract and more money without even asking it.

Client 2. Selling aviation books and online courses through Woocommerce - Worldwide. More then 20GB of data in shop. Blazing fast. No problem whatsoever. TTFB below 500ms.

Another case. Server CX31, running Onyx (still not upgraded) and more then 70 wordpress installations yes with many stores. All flies. Yes we done installation on all our clients websites because we have full service, but we didn't do anything special we just used most quality themes from Themeforest (big selling number is indication of good theme and constant development) and we made sure that Wordpress is installed by standards.

Btw yes all our servers are running same OS CentOS Linux 7.7.1908 (we plan to upgrade to new centos next year when Obsidian will support it)

To focus on OP question. What did we do, did we run some special trick? Some utterly magical trick? Answer is NO! Everything work for us in Plesk out of the box. Apart from usual server config and security.

The only difference i see here is that we rely on FPM served by Apache.

I would like to suggest that Caching is essential but reason for your low score can be anything. It can be bad theme, bad plugin. You really need to dig it out. These days it's rather simple. You have online benchmark tools, you have amazing dev. tools in Firefox and especially Chrome where you can see what is wrong.

W3TC is amazing plugin on it's own we use it almost exclusively but it won't work well with every theme. Especially minification part. People think they install plugin, tick few options and their score should be magically improved. It does not work like that. OP mentioned he is using gzip in W3TC. Well... One can enable it in W3tc but in reality it may not work at all. Have you actually tried to run W3TC system report and see is everything needed for caching actually available on your server? System report in W3tc will spit out red and green results.

To finish this up - Wordpress started as a blog yes but please stop spreading misconceptions. People started in belief that Earth is center of universe. Today we know better. OP should focus and measure what is causing his issues. At this point i am pointing out themes or bad wordpress setup before going to dig server and make potentially harmful changes.

Or before i suggest him to change his complete framework or workflow.

Merry Christmas to everyone.

 

[jojo]

I recently switched over to another hosting provider and the performance did improve a lot, still not to the point I'm aiming for, and I think this is mostly due to bad nginx caching setup (cookies).

 

[Martin.H]

you are spreading wrong information.

No offensive, but nothing I said is wrong! If so proof it is wrong, easy to say someone is wrong. And especially tell me whats "wrong".

This is just my opinion, you do have a different one, I respect that, but you do not seem to respect my opinion and thats why you want to silence me. I dont like that.

You are speaking about a good TTFB, but hey a TTFB of

below 500ms.

(what exactly under 500ms????) is very very poor.. specially when it comes to selling and native SEO (without SEA!)
My shop does have a TTFB of 89ms (Average Worldwide!!!)

Googles recommendation is 200ms, perfekt would be 100ms.
Everything above 600ms (Source) Google will not meassure anymore

I feel like your one of the "Wordpress Guys" whos answer to every problem/request is "Wordpress" yes Wordpress moved/eveloped a lot, but in the end, everytime I use (have to use) it I notice its still not made for really good commercial use.

GTmetrix gives 79% (but that's relative anyway), Pingdom gives 87%.

In the end, OP asked about TTFB not optimization of the whole website. TTFB is just a part of SEO, GTmetrix/Pingdome and Google PageSpeed Insights are meassuring the whole SEO value, not just TTFB, so these values are mistaking here. Especially if we dont know if they apply to your Mobile, or Desktop site? As far as I know GTmetrix and Pingdome does not meassure mobile SEO scores and this is the first problem with them. As Mobile SEO is quioet different then the one of Desktop sites!

Just use GPSI if you want to rank good in Google, and use others just in addition for this, but not as primary (technical) SEO Analyser.
Just my site, it get 100% on GPSI (Google PageSpeed Insights) Mobile:

This is mobile, Desktop nearly every Page scroes 100%, thats why you do not mention it anyway, as its to easy.
Again: infos like this, are normally offtopic, as they do not have anything to do with the TTFB in first place.
In the end the Plesk standard config is good, but seriously, its still a standard config not optimized. So my advise: optimize it for your Server to fit your /Setup/CMS/Usecase to 100%.

Also: if you just dont know better and you are happy with a TTFB of "under 500ms" well go with it, but pls do tell others this is fast/special. Its not

Client 1

Client 2

A lot of talk, without proof.. also like I said: if you are happy with Wordpress thats ok.
Dont get why you have to go bananas, becasue I said that Wordpress is not made for professional business. Ofc I have seen some good pages made with Wordpress and even some big stores, but I'm pretty sure they would have done better with other CMS. If you dont think so I'm again ok with that but dont try to silence people because you/they do have a different opinion!

Choosing Wordpress for me just makes sense when money is very short.
This means you are not aiming for "quiality"/"time" but for "investing less money" and thats ok.
But if you buy a cheap car, dont expect it to be as fast/nice as a expensive one.

Would advise looking at that triangle:

 

[seqoi]

No offensive, but nothing I said is wrong! If so proof it is wrong, easy to say someone is wrong. And especially tell me whats "wrong".

From the drama you posted i can see that you have zero experience in real life scenarios. So basically you said i was lying and i need diploma work posted here to convince you that i am speaking truth. With proofs. Gimme a break. If i were to post screenshot and proof every time internet nobody ask me to - i wouldn't had time to get real work done or to live actually.

But at the same time OP actually posted what i was basically reported. Which was (to remind you) either server issue or WP setup issue. As you can see it was server issue.

So i will repeat: instead of you spreading wrong information and suggesting OP to change complete workflow and migrate to other platform (which was suggestion based on your lack of experience - hence it's wrong) - better avoid thread and keep on with your life. Learn a bit. That's what i do. Learn something new every day or week.

I'll mark your response for myself. It's good candidate for best biscuit of 2019.

Happy new year.

 

[Martin.H]

hence it's wrong

Again: if your opinion is that I'm wrong I have to accept that. But please do not state I'm wrong and then just talking ****.

There is a big difference between "I think you are wrong", "I have a different opinion" & "you are wrong"
Wrong or right is not absolut. Pls get that.

instead of you spreading wrong information

You again not said what was wrong.. you just talk and talk and talk, without saying anything but "wordpress ist best". I accept you having a different opinion, but I do not accept you calling mine "wrong" without any proof.

You are pretenting your OPINION as only true thing. Please lear that opinions can vary. Calling another opinion wrong without proofing it is just stupid.

But you will may not get it.. probably never.

#Edit: I will not respond to this Thread anymore. Ty for understanding.

 

[seqoi]

"wordpress ist best". I accept you having a different opinion, but I do not accept you calling mine "wrong" without any proof.

You are pretenting your OPINION as only true thing. Please lear that opinions can vary. Calling another opinion wrong without proofing it is just stupid.

But you will may not get it.. probably never.

#Edit: I will not respond to this Thread anymore. Ty for understanding.

No problem i understand. That's ok if you want to twist your reality to accommodate your distorted view i am fine with that.

However you spread even more false information. This time even lies. I NEVER said that wordpress is best. You may stop doing that or i will report you.

"Calling another opinion wrong without proofing it is just stupid." - Dear friend your logic doesn't mate any sense.

You do realize that today even in 2020 there is thousands of people which think earth is flat. I can say to them that their opinion is wrong without providing them truth. Why? Because other people provided that already. And there's common sense to it if you observe reality around you.

You keep continue asking me for proof yet you are the one which spitted out false argument about CMS platform - without ZERO proof - which is why i reacted here. How convenient.

Same in this thread. Original poster already provided you with the information that in the end it was server side issue - which is what i suggested. Compared to your switch COMPLETE environment suggestion - hence very wrong and false). Yet you continue to beat this dead horse, you posted few cartoonish graphs and you think you proved something.

Even after original poster himself told you what was the reason. Can you realize what is wrong here really?

You even said "google said" which is laughable. Google said something about something and then 10 years later their own claims are being changed, or even deceased.

After you said TTFB of 500ms is poor and when you asked me what is below 500ms, right there i could said you had zero experience on running larger scale projects. Running one website, with one value does not make you expert to suggest to someone how they should change their complete environment. Simple as that.

Out of fun i just measured Microsoft News portal TTFB which is running wordpress. Stories | Microsoft news, features, events, and press materials - their TTFB is around 300ms. I just measured. Here waste time yourself - Byte Check - Check Your Time To First Byte - By your logic you would think that Microsoft would do their best to make it under 89 right...well..no...Why? Because in practice what you said of course doesn't make any sense.

Once again happy new year. Live long and prosper.