WPMUDev heavy toll on server

Greetings,
Since yesterday I decided to use your system to update, secure and optimize all my websites via the WPMUDev HUB and the plugins Defender, Snapshot, Hummingbird and Smart Crawl.

I got rid of my previous service to do mass maintenance (CMS Commander), connected WPMUDev on all my websites and set up all the plugins...

The problem is, since yesterday, my server is dying, because since I installed all your plugins (just the 4 I mentionned + WPMUDev hub + sometimes Smush it) my CPU and RAM is saturated...

So buying more CPU and RAM is of course an option, but first my host advised me to be clear on which plugin exactly was taking all the ressources...

And also, I'm wondering if it's normal that your plugins (especially some dedicated to wordpress OPTIMISATION) are taking such a heavy toll on my server's ressources ?

I mean, I'm not a giant, I'm managing only 19 websites, and if I have quite a few plugins on them, I'm not going crazy overboard either... Also I'm positive that the WPMUDev plugins are the cause, because I didn't install anything else (if anything, I removed some plugins that were doing the same jobs as WPMUDev plugins).

I have to say I'm a bit surprised and disappointed : I trust you, so I used your plugins first, but I never expected they would take such a heavy toll on my server and would require me either to remove them (which is a LOT of time lost), or to pay double to get a more powerful cloud (I'm currently using a 2CPU-6GoRam cloud hosting from Infomaniak).

Please let me know what you think, and which main plugin is taking so much ressources, or if there is a way to still use your tools without needing to get a more powerful server... Thanks

  • James Morris

    Hello DJExp

    I hope you are well today.

    Hmm... This seems rather odd. I have a small cloud server with the following specs.

    CPU: 1 vCore
    RAM: 1 GB
    SSD: 50 GB
    Architecture: VMware virtualization technology
    Processor: Intel® Xeon® Processors: Intel® Xeon® E5-2660 v2 and Intel® Xeon® E5-2683 v3

    With this setup, I've tested member's sites that measure in the GB of size for images and hundreds of MB in size for databases and I've yet to max out the CPU. At worst, I've capped out the RAM, but it was during very heavy database operations.

    So, I'm wondering what is specific to your configuration that may be causing such a massive spike in CPU.

    I will be more than happy to dig into this further for you, but I will need access to your site and server as well as a bit more information about when you were seeing these spikes.

    Can you tell me (roughly) when you saw the spikes? Were you performing Bulk Smushing? Defender Scans? Snapshot Backups?

    Of these, Snapshot is the most expensive on CPU and RAM. This is because it is a software based backup solution that uses PHP to perform the backups. We are currently working on a full rewrite of this plugin that will allow it to leverage system libraries (read native system commands) to perform backups, which is MUCH less expensive on CPU and RAM.

    Defender can be a little expensive on CPU initially while it loads and scans the file system. This generally is not a problem and does not happen frequently. So, it really should not cause that great of an impact on your CPU load.

    If you would like me to take a deeper look at your configuration to determine if there's anything specific that might be causing these symptoms, would you please grant me access to your site and server so I can perform some advanced debugging?

    Please visit the Contact page and complete the form with the following information:

    https://premium.wpmudev.org/contact/#i-have-a-different-question

    Option: I have a different Question

    Subject: "Attn: James Morris"

    In the Message box, please provide the following:

    - link back to this thread for reference
    - any other relevant urls

    - Admin login:
    Admin username
    Admin password
    Login url

    - Hosting Control Panel Login
    Admin username
    Admin password
    Login url

    ~OR~

    - FTP credentials
    host
    username
    password
    (and port if required)

    Best regards,

    James Morris

  • DJExp

    Hi James,

    The support of my host told me your plugins were the most likely culprint when I told them they're doing caching, optimization, backup, securities, etc...

    Also, the problem started to appear after I massively installed your 4 plugins (Defender, Snapshot, Hummingbird and Smart Crawl + the WPMUDev dashboard plugin) on all my websites.

    Now I deactivated Hummingbird and Smartcrawl everywhere and the heavy toll on CPU and RAM seems the same... So now I should test deactivating also all remaining WPMUDev plugin, and if it doesn't change the impact on server significantly, it means your plugins aren't reponsible and the hypothesis was wrong...

    I can do that, but I have a question first :

    IF I deactivate plugins such as Defender, Snapshots, Hummingbird, etc... not delete, just deactivate), and then re-activate them again... will I need to go through all the setup again (the .htaccess stuff, the tests, the optimizations, the snapshots creation, the scans, etc.) ?

    If you can confirm that doing that test and deactivating the remaining plugins (Defender and Snapshorts + the WPMUDev dashboard) on all my site will not result in me having to reconfigure everything (which took me HOURS), I can do that.

    That way, if I remove EVERYTHING WPMUDev related and the problem still exist, it means it's Infomaniak's problem (my host) and they cannot blame you anymore...

    With all that said, I'd still be glad if you could have a deeper look at my configuration...

    So to answer your question :

    Can you tell me (roughly) when you saw the spikes? Were you performing Bulk Smushing? Defender Scans? Snapshot Backups?

    I noticed the problem yesterday, but it seems to have started a bit before that. Please have a look at the CPU and RAM monitoring stats attached.

    It seems that CPU problems started 2 days ago (March 21), and RAM problems started beginning of last week (March 12-13). I have to say it match the time I started putting all my website maintenance's task with your plugins...

    Now as you can see on the graphs, there isn't really "spikes", it's more like a plateau, a general saturation.

    I'll do the process to grand you the access now.

    Thanks a LOT for your help and for looking into it ! Thanks !

  • James Morris

    IF I deactivate plugins such as Defender, Snapshots, Hummingbird, etc... not delete, just deactivate), and then re-activate them again... will I need to go through all the setup again (the .htaccess stuff, the tests, the optimizations, the snapshots creation, the scans, etc.) ?

    Deactivating them should not revert any of the settings, as long as you don't click "Delete". However, there's actually a very quick and easy way to test plugin effects and guarantee that no settings are not removed.

    The simplest way to perform a plugin conflict test:
    * Login to your site via FTP
    * Navigate to wp-content
    * Rename the plugins folder to _plugins
    * Create a new folder named plugins
    * Move the plugin(s) folder (ex: membership) you want to test from _plugins to plugins
    * Then, one at a time, move other plugin folders over from _plugins to plugins until the symptom returns

    NOTE: While you are doing this, do not visit WP Admin -> Plugins on your site. This will ensure that your plugins remain active and all settings will remain when you copy them back over.

    There is a pretty well known problem with WordPress relating to the native Heartbeat control. This is particularly noticeable with plugins (and themes) that have a lot of front-end scripts. The vast majority of our plugins do not have many front-end scripts, but performing these steps can help just about any installation reduce server load.

    The following article from SiteGround has some excellent information on this issue as well as some steps to resolve it. https://www.siteground.com/tutorials/wordpress/limit-heartbeat/

    You can learn more about the Heartbeat functionality from the WordPress Docs:

    https://developer.wordpress.org/plugins/javascript/heartbeat-api/
    https://codex.wordpress.org/Function_Reference/wp_heartbeat_settings

    I hope this clarifies a bit.

    Best regards,

    James Morris

  • DJExp

    Hi James,

    So I got words from my host's support. It seems that the culprit wasn't the WPMUDev maintenance plugins (first hypothesis). It seems the file "wp-admin/admin-ajax.php?action=wp_lp_installer&nonce=421c" received a crazy amount of "post request"... The blocked the adress and now my server can breathe again.

    What's interesting is that this file seems to be related to the Heartbeat functionnality you just told me about, so I have to say I admire your guts feeling ! :slight_smile:

    Anyway, now I just have to find a way to secure that website (I'll start with your Defender plugin !) to prevent those attacks on this file, and I'll probably deactivate "Heartbeat" as well, and this should be good to go.

    Thanks a lot for your help in this !

    Dushan

Thank NAME, for their help.

Let NAME know exactly why they deserved these points.

Gift a custom amount of points.