Page broken on mobile

I have quite the strange issue I have trouble with debugging. One of our pages just stopped working in two ways in its mobile version: the front page includes a slider made with “Smart Slider 3” which doesn’t load anymore, and the mobile navigation bar no longer can be used (no idea if that’s part of the theme, but if relevant, our theme is Customify). But accessing the page from a desktop computer (and funnily enough from an iPad too, but not from any mobile browser on a smartphone, even with desktop mode enabled, though Safari on iOS wasn’t tested) doesn’t show any issues whatsoever. What makes this all the more strange is, we have another website on the same server with the exact same theme, though with different plugins (e.g. Smart Slider isn’t used). It doesn’t display any issues.

So first thing I tested was if Smart Slider was the cause, but deactivating it doesn’t fix the page. Even the nginx log didn’t say anything. So next I enabled debugging:

define( 'WP_DEBUG', true );
define( 'WP_DEBUG_LOG', 'wp-content/debug.log' );
define( 'WP_DEBUG_DISPLAY', true );
define( 'SCRIPT_DEBUG', true );
define( 'SAVEQUERIES', false );

but the only message the debug.log shows me over and over is

[10-Oct-2024 11:48:55 UTC] PHP Deprecated:  rtrim(): Passing null to parameter #1 ($string) of type string is deprecated in /var/www/wordpress/<page name>/wp-includes/formatting.php on line 2819

which I hope isn’t the issue, as that’s something WordPress would need to fix. But at least now I get a new message in nginx’ error log. But it’s only lines from open() and FastCGI complaining that some pictures are missing. But these pictures have been deleted months ago and shouldn’t be referenced anywhere anymore. Still, I can see them referenced in the mysql database, though I can’t tell in which table (found them by searching a mysqldump of the database with ripgrep-all).

The only thing that may be related is this message from nginx’ access.log

134.130.104.122 - - [10/Oct/2024:14:06:30 +0200] "POST /wp-admin/admin-ajax.php?action=as_async_request_queue_runner&nonce=9671aac4fc HTTP/1.1" 499 0 "-" "WordPress/6.6.2; https://domain.tld"

But HTTP error 499 is about the client prematurely closing the connection. As this happens with both Chrome and Firefox for Android, it’s unlikely that’s the cause.

But aside from being a slide annoyance, that shouldn’t be a cause for this behavior. Other than that, nothing. And there are also no debug messages that are being injected into the pages HTML. Even the php8.2-fpm.log doesn’t contain any messages from today.

So, how can I figure out what’s wrong to either fix this or at least open a relevant bug report in the right place? As this only happens on phones and I don’t know of a mobile browser capable of displaying a developer console to show some error messages, and since I can’t produce any logs on the server that contain any relevant messages, I have no clue what could be the cause.

Our setup: Debian Stable 12.7 with php 8.2.24 and nginx 1.26.2, WordPress 6.6.2 on all websites and all plugins and themes up to date.