Interpreting pageload results - very slow!

[bartman9]

My vBulletin based site (running on IIS) is very slow and I need some help interpreting the pageload results. Here are the results:

http://www.webpagetest.org/result/140603_DX_KRZ/

The time to first byte is very slow. Is this a server issue? I just switched the site over to a dedicated server and I thought it would speed things up. But it is even slower now.

Also, the cron.php seems to be the next bottle neck. I don't know anything about php code, so any help here is greatly appreciated.

[RichieBoy67]

Did you configure and tweak the server? You need to tweak your my.cnf, php.ini, etc..

What do your server logs say? What cron task is using so much of your servers resources?

[Max Taxable]

Fully 50% of your pageload is custom font files. Most browsers ignore the fancy fonts and display their set default, but the files still download to the registry. Most visitors will never see them, they're not worth the bandwidth they consume. And most especially, not worth nearly half of your page size in kilobytes.

Try the same test using vB default style, with all hooks off and observe the difference it makes. First Byte Time will be better as well. You are loading more than one megabyte on browsers. Not as bad as many I've seen, but still a bit fat nonetheless. Show me a fat page and I'll show you a slow page, every time.

Also, we have a Optimization Workshop Here with plenty of real-life case studies of problems such as this, solved.

I use WPT with IE10, exclusively for the full picture:

http://www.webpagetest.org/result/140604_JM_RW8/ <--- The CMS

The forum is a little more raunchy for pageload: http://www.webpagetest.org/result/140604_RW_S3B/

[Zachery]

Originally Posted by Max Taxable

Fully 50% of your pageload is custom font files. Most browsers ignore the fancy fonts and display their set default, but the files still download to the registry. Most visitors will never see them, they're not worth the bandwidth they consume. And most especially, not worth nearly half of your page size in kilobytes.

Try the same test using vB default style, with all hooks off and observe the difference it makes. First Byte Time will be better as well. You are loading more than one megabyte on browsers. Not as bad as many I've seen, but still a bit fat nonetheless. Show me a fat page and I'll show you a slow page, every time.

Also, we have a Optimization Workshop Here with plenty of real-life case studies of problems such as this, solved.

I use WPT with IE10, exclusively for the full picture:

http://www.webpagetest.org/result/140604_JM_RW8/ <--- The CMS

The forum is a little more raunchy for pageload: http://www.webpagetest.org/result/140604_RW_S3B/

There is nothing wrong with webfonts, as long as you're using common, or Highly optimized fonts. my style iconic uses an optimized webfont.

You want to use something like OpenSans from google's font api, which means that there is a very good chance that the user already has this document in their browser cache. Its also fairly tiny.

[Max Taxable]

Originally Posted by Zachery

There is nothing wrong with webfonts, as long as you're using common, or Highly optimized fonts. my style iconic uses an optimized webfont.

You want to use something like OpenSans from google's font api, which means that there is a very good chance that the user already has this document in their browser cache. Its also fairly tiny.

If they comprise half of a 1 megabyte pageload, there's ALOT wrong. In this case they're not tiny, they are 50% of the entire page load. For stuff 80% of people won't even see.

If font files are 2% of the pageload or less, it's not a issue. This is 50. That's half.

[RichieBoy67]

Originally Posted by Max Taxable

If they comprise half of a 1 megabyte pageload, there's ALOT wrong. In this case they're not tiny, they are 50% of the entire page load. For stuff 80% of people won't even see.

If font files are 2% of the pageload or less, it's not a issue. This is 50. That's half.

Yeah but I do not believe it is fonts that are causing the browser to take that much time... It is pretty clear to me that he has a server issue or something else going on beyond fonts.

--------------- Added [DATE]1401919701[/DATE] at [TIME]1401919701[/TIME] ---------------

## First View is 21.242s! Mine was 6 and I thought it was slow. This is more than fonts.

[Max Taxable]

Originally Posted by RichieBoy67

Yeah but I do not believe it is fonts that are causing the browser to take that much time... It is pretty clear to me that he has a server issue or something else going on beyond fonts. .

If we pretend half a megabyte of data just loads without using any time, yeah.

The idea is to get rid of this question, when I said this:

Try the same test using vB default style, with all hooks off and observe the difference it makes.

[RichieBoy67]

Originally Posted by Max Taxable

If we pretend half a megabyte of data just loads without using any time, yeah.

The idea is to get rid of this question, when I said this:

I'm not really disagreeing you brother. While a half a meg is nothing by itself when you put a ton of them together you are looking at problems.

In regards to this slow down we really do not have enough information to troubleshoot this.

Is it slow all the time or just when a certain cron task is running?

Has anything else been changed besides moving it from one server to another?

If he is on a new dedicated server though and he just moved his site over from another one then my opinion would be the server or something related to the move.
If it was a plug in or theme it would most likely would have happened on the other server as well.

Many people grab a root server and do not realize it has to be configured.

[Max Taxable]

Originally Posted by RichieBoy67

I'm not really disagreeing you brother. While a half a meg is nothing by itself when you put a ton of them together you are looking at problems.

In regards to this slow down we really do not have enough information to troubleshoot this.

Is it slow all the time or just when a certain cron task is running?

Has anything else been changed besides moving it from one server to another?

If he is on a new dedicated server though and he just moved his site over from another one then my opinion would be the server or something related to the move.
If it was a plug in or theme it would most likely would have happened on the other server as well.

Many people grab a root server and do not realize it has to be configured.

This is why the basic troubleshooting task of testing with hooks off and with default style needs done.

Likely he had these same issues with his previous host - he as much as said so, the switch to dedicated hosting was the speedup hoped for.

People get trapped into thinking CDNs or dedicated hosting are magic bullets for performance. They are not. Optimization, is.

[bartman9]

Thanks for talking this out guys, I will try to shed some light on this.

- The website was slow on the shared server too. That is why I decided to move to the dedicated server, thinking it would be much faster. I am still using a shared mySQL database.

- I disabled all hooks and the website loaded much faster. See results:
http://www.webpagetest.org/result/14...KMG/1/details/

So when I disabled these features (in config.php), is that turning off all plugins and/or disabling fancy fonts?

Knowing this, what should be my next step? I am not much of a PHP guy and I really don't know much about configuring servers. All help is greatly appreciated!