One of my friends recently asked me a question whether using too many plugins made any difference on how his blog worked and performed, I thought it best to give that answer to a broader audience, so here goes.
Does It Matter If You Use 1 or 100s of Plugins?
The answer is yes, it does matter, each and every plugin you use has to be processed before an actual page is generated, so all the plugin files are compiled in order to get the intended output.
This may in the end not actually make a big difference in your blogs load time, however it may eventually slow it down a little bit.
Do Plugins Add Overhead To Your Site?
Yes at times they do add overhead to your site, for example quite recently I was using a plugin that added 180Kbs of JavaScript code to each and every page on the site, however that plugins functionality was only useful on a single page.
You should periodically check these things out as you may be unnecessarily increasing your page load times, if possible try and find an alternative or contact the plugin author to add an option to limit overheads to only those pages where the plugin functions.
Should I Stop Using Too Many Plugins Then?
Not exactly, like I said it does affect your blog load times and you might want to take a look at your plugins to see which of them are really useful and maybe remove some of them.
I usually change several things on the site quite often, and with that I stop using the functionality provided by certain plugins. I run a weekly or bi-weekly maintenance or spring cleaning of the blog, in which I take a look at all the active plugins and see whether I am using them or not, and disable those that I no longer use.
Many a times you will find that there are plugins that are unused and are still active, it is best to disable those plugins while you are not using them and enable them back whenever you need to use their functionality.
What Type Of Plugins Should I Avoid?
There is no real rule-of-thumb for me to say what to avoid and what to use, each plugin’s functionality is different and useful, however something that is useful to me may not be useful to someone else and vice versa.
However as a general rule, on a high traffic blog try and avoid plugins that make many database calls, since it may affect performance, it is ok to make around 10-20 queries per page, however anything more than that degrades performance and may crash the DB on high traffic sites.
What Else? Tell Us Your Tips
There may be several other things you might want to add, go ahead and tell us what you think should be ideal when it comes to using plugins, we would love to hear your opinions.
Also, I will be doing a follow up post addressing several things that have been left unanswered in this post, if you have any questions feel free to leave a comment, I will try and answer them in the follow up post.
Avoid plugins that don’t get updated regularly. As WP get upgraded each time, some plugins may not be compatible.
“10-20 queries” is completely arbitrary. Anyone knowledgeable enough to open a plugin and count the query calls should know how beefy a machine they need to handle the load. Don’t skimp on plugins if it improves the experience for your site’s visitors or makes your life easier managing things. Even on an underpowered box, eAccelerator, a caching plugin, and compacting & compressing your javascript/css will lend far better performance benefits than just dropping a few plugins.
That being said, many plugins are terrible when it comes to managing javascript. Make sure they make use of ‘wp_enque_script’. I’ve seen sites that load 2 separate lightboxes & 3 copies of prototype & jQuery on each page just because the plugin authors & site owners didn’t know better.
@Richard I meant 10-20 queries on the whole for the page, not a single plugin, I think there is a way to determine how many DB hits each page performs for the output to be shown.
Is there a way to know which tables each plugin adds to my mysql database for my site so after deactivating, deleting the plugin I can cleanup the database.
This plugin catches many of the leftover, orphaned options that were added to wp_options by plugins and themes that have been deactivated and uninstalled. It also allows you to search Google for references to the data it finds.
Always backup your database before deleting any of the orphaned options the plugin finds.
@gestroud
Which plugin?
Note to self: Never post at 3 a.m. :-S
http://wordpress.org/extend/plugins/clean-options/
One of the things I suggested to WP after the 2.8 release is to put a field in the plugins for a date that it was installed. That way, I can visually do a quick check to see if it is a plugin that I need to disable because the author doesn’t update it regularly. If anyone thinks that’s a good idea too, maybe you can go to WP and look for my submission by my user name webduck
When I analyzed my blog’s traffic, I found a substantial bandwidth was used by prototype.js which is a huge file. It was called by a plugin though WorPress core doesn’t use it. I deactivated it immediately.
I feel many of the plugins add only a fancy value except a few core ones like Akismet, Google Analytics, Lester Chan’s plugin suite, and a few SEO enhancements.
As Keith says, you have to analyze the mischief of the plugins individually and there is no shortcut solution.
My dos pesos!
S.K
I try to keep my plugins to 20 or less now, remove the ones that I don’t really need and keep the basics.
I have a similar query to Miroslav – how do I know which plug ins are making over 20 database queries per page? Sorry if that’s a dumb question but I’m curious. Thanks for an informative post.
This post is pretty much all wrong.
The number of plugins you have activated essentially makes no difference in your blog’s performance. It’s just a simple
include(). This is proven by that if you merge all of your plugins into a single plugin file, you will notice no increase in performance.However, what plugins you have activated does matter.
Having 100 plugins that are just a filter or something simple is much better than a single plugin that calculates the first million digits of pi.
In short, it’s quality not quantity that matters. Don’t have any plugins you don’t actually need and try to use popular plugins as they tend to be better coded (not always, but often).
@Viper
Of course it does matter. Every code loaded/interpreted/to be compiled/etc. needs time to process. Every software that needs time to show its features/functionality performs well or it doesn’t. So yes it can influence your performance. As programmer you should know.
“Having 100 plugins that are just a filter or something simple is much better than a single plugin that calculates the first million digits of pi.”
Well, what would be more sophisticated to determine? The first million digits of pi, or (for example) show statistics of your weblog? To show your statistics would have lots of more code then just determining all digits of pi, but determining would have more impact on your performance.
You’re missing my point.
My point was the complexity of the plugins and what they do matters, not the mere number of activated plugins. Having one plugin activated is not always better than having 100 activated.
@Viper007Bond If you noticed here is what I actually said
“This may in the end not actually make a big difference in your blogs load time, however it may eventually slow it down a little bit.”
I never said 1 or 100 plugins slowed down things, I just said it makes a difference, compiling 1 file vs 100 files takes a performance hit, if you think that clubbing 100 files into 1 file makes a difference I would really say it does not, because eventually you are still going through the same compilation, compiling 10 lines in a file and 10000 lines in a file is not the same, just that you would save a bit of time loading different files.
I would avoid automatically updating. I just spent two hours fixing a plugin after update that destroyed the plugin.
Using too many plugin is the reason I created GD Press Tools plugin. With this one plugin you can replace over 20 plugins that do simple things like adding columns to grids, working with posts and many more things.
So, instead of loading 10 or 15 plugins, WordPress will load only one. And this way it’s easier to find out if something goes wrong, and you don’t need to go through each plugin finding the problem.
See my above comment. 10 or 15 plugins is no slower than one big plugin. It’s all about what the plugins do rather than the number of them.
And see my comment below that one.
This is so really really –not– true. You simply cannot say so before it having benchmarked.
Definitly it has negative impact on WP. I will release new version of GD Press Tools tomorrow, and it features memory measurment during the forming of the page in admin section, and you can see how much memory is used.
Load any admin page and note the memory used. Disable all plugins (leave GD Press Tools only), and see how much memory WP uses now. Right now, my blog has 16 plugins. With all 16 dashboard takes 31MB, with only GD Press Tools it uses only 24M. Similar values are on the actual blog pages also.
@Milan do you have any performance statistics for all the plugins you combined for GD Press, it would be interesting to see how it does when you use all 15 vs GD Press.
It would be great to see that. I will try to set up a test blog and to gather as much of the smaller plugins GD Press Tools replaces and than measure used memory and even page load speed and some other things. When I do that, I will post results of analysis on my Dev4Press website and I will notify WeblogTools.
I am very curious about the very issue that Viper007Bond brings up. If I combine 100 plugins into a single plugin — adding all the exact same functionality, hooks, filters, etc. — will it make *any* difference? I look forward to the results of Milan’s tests, which seems to me to be the right way to test this issue.
Using Firebug, you can test the impact of each plugin’s speed on your site. I’ve got almost 30 plugins running and found my site slowing to a crawl over time. By enabling plugins one by one and timing the reload speed of the page, I could find which ones were slowing me down.
For more information, see Speed up WordPress by Disabling Plugins
I try to keep number of plug-ins less than 15 on my blogs. Most importantly I ask friends to avoid plug-ins that do multiple database queries, or complex full text searches. I often use YSlow (and now google page speed) to check if blog is becoming slow. Actually a lot of plug-ins are fine unless you hit first page on Digg 😉
Here’s a couple of suggestions:
1. Some people continue to use Plugins that they’ve used through many iterations of WP, that are no longer necessary due to core enhancements and widgets provided with newer version, so you should ALWAYS review what plugins you use to see if they are still needed when upgrading WP;
2. SOME plugins can very easily be replaced by simply adding some lines of code to your functions.php file, or by modifying your Theme’s header and/or other template files (ALWAYS backup your theme’s functions and template files before modifying)…..I realize that can be daunting for some, but it’s easier than you may think and work trying…..THIS METHOD allows you to “customize” which pages actually need to have that code and which don’t, avoiding code overload on pages that don’t. It’s worth the time to see exactly what the plugin does and see if you can replicate it’s function without installing and activating it.
3. SOME plugins add some very pretty, fancy ajax-y features that look nifty, but can slow down your page load. Ask yourself if those features really benefit your readers and are really necessary? If so you may consider upgrading to a Premium Theme that offers those features without the plugins, or have someone custom develop a theme for you – if you ask your readers, very few of them would choose prettiness over speed at getting what they want to read on your site.
Here’s a request for Plugin Developers:
PLEASE if you’re adding a lot of lines of code to every page that may not be necessary (PodPress is a good example – only maybe 1 in 20 of my posts have a Podcast but PodPress adds code to every page)…..add an Option ( a tick-box maybe? ) for blog owners to check if the page NEEDS your code or not…..
@Keith – this is a great discussion to be having – there are so many plugins these days, many of them are really creative and well done, and it’s tempting to want to use so many of them…..it’s a good reminder to everyone that they should consider the effect on their site before loading up on all the cool plugins….
Yeah, I believe a post that really outlined what us mere mortal WordPressers could do, step by step to check things on our individual installs would be immensely popular. Matthew points out Firebug, which I’ve used for over all page performance tuning, but I did that following a tutorial. Having something similar for WP and ones use of plugins (themes, too?) could possibly become one of the “indispensable” tools that ever WP admin uses.
Just saying 😉
Placing plugins that are not actualy used over the online version of WP means a using a lot of resources for no reason. Heryby if a use want to test a plug-in before uploading, he/she can turn their PC onto “local server” and try whatever they want to after being satisfied with the results.
Well, to many plugins aren’t good. However, if the only problem is the vast amount of SQL queries eating a large bandwidth between the data server and your application server, you can use cache plugins to resolve this problem (partially or totally).
Another great question is whether it’s better to use a few complex plugins or a whole set of small plugins? Which case is better?
In my opinion, a few complex plugin performs better, because of the less amount of function calls (hooks, callbacks, action, auxiliary functions, etc.).
You simply cannot say so, you just need to test and benchmark it to prove it. It’s so easy talking saying less is better, or much/many is worse. Every code is still translated to assembly. It may seem a simple function call, but in assembly it still can be a hell of list of code.
I recently started using the PHP Speedy plugin which combines all CSS and Java files into one. With a tiny bit of code you can also make this load in your footer which is supposed to be a better practice as well. This way all your content displays on screen even while the Java and other stuff is being loaded due to other plugins.
I tried to cut down from 28 plugins to 10, and it didnt effect speed at all.
I have a question I’ve always wondered about…
If you use the widget system in wordpress, is there a way
to put in custom sidebar code?
like just as an example, the post here about displaying recent comments without using a plugin or widget..
IF you use the widget system, instead of sidebar.php, then how can that be possible?
This question pertains to why I’ve never used the widget system until very recently.
Whilst not directly related, I use a great plugin called TS Custom Widgets (http://www.thaslayer.com/free-plugins/) to enable and disable widgets on different parts of my blog. Lets your blog have a bit more of a dynamic feel, and also allows you to tune out widgets on high load pages.
I have published first in the series of WordPress benchmarking articles, so check it out: http://www.dev4press.com/2009/.....istration/