Hi,
Thank you for the improvements. I appreciate this has been taken care of
From my understanding, at least within the WordPress environment, autoloading is beneficial when the data is required on every page load or on key pages (e.g., homepage, SEO-focused content, etc.). However, autoloaded data is loaded on every request — including backend pages, AJAX calls, cron jobs, and other processes — whether it is needed or not.
As a result, the more autoloaded options there are, the greater the impact on TTFB and the higher the memory usage per request.
In short, while it may be tempting to autoload many options for convenience, this is often not ideal. Each website has different needs and should carefully evaluate what truly needs to be autoloaded. WordPress generally recommends keeping autoloaded options under 800 KB.
I reviewed my wp_options table and noticed multiple WPML (and add-on related) entries set to autoload (or eligible for autoloading). I haven’t updated to the latest version yet, and I understand that this may already address part of the issue. However, if the only recent change concerns wp_installer_settings, I believe there may still be room for optimization.
In my case, there are 85 WPML-related autoloaded entries. At first glance, some of them seem difficult to justify as autoloaded, for example:
- _wpml_jobs_not_notified
- wpml_strings_need_links_fixed
- otgs-installer-log
- wpml-updates-log
- _transient_wpml_admin_text_import***
While some of these options are small in size and individually, have a small impact, the broader concern is cumulative impact. If each plugin does not carefully manage autoloaded options, this can significantly affect overall site performance.
Can you explain if, and why these options are required?
In my situation, I am approaching the recommended 800 KB threshold. For websites that use WPML more extensively, or many plugins, optimizing autoloaded options could have a meaningful performance impact.
Thanks!