Bruno Cantuaria

Additional note:

Here’s an entry from the Purge Log:

04/26/24 15:41:35.969 [**masked** 1 Kgp] 💓 ------POST HTTP/1.1 (HTTPS) /wp-json/wp/v2/posts/601385
04/26/24 15:41:35.969 [**masked** 1 Kgp] Query String: _locale=user
04/26/24 15:41:35.969 [**masked** 1 Kgp] HTTP_REFERER: https://www.**masked**/wp-admin/post.php?post=601385&action=edit
04/26/24 15:41:35.969 [**masked** 1 Kgp] User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/124.0.0.0 Safari/537.36
04/26/24 15:41:35.969 [**masked** 1 Kgp] Accept: application/json, */*;q=0.1
04/26/24 15:41:35.969 [**masked** 1 Kgp] Accept Encoding: gzip, br
04/26/24 15:41:35.969 [**masked** 1 Kgp] Cookie _lscache_vary: admin_bar:1;guest_mode:1;logged-in:1;role:50
04/26/24 15:41:35.969 [**masked** 1 Kgp] X-LSCACHE: true
04/26/24 15:41:35.969 [**masked** 1 Kgp] X-LiteSpeed-Purge: public,stale,482_Po.601385,482_URL./**masked**/,482_Po.601376,482_Po.601391,482_T.1,482_T.41016,482_FD,482_A.103,482_F,482_H,482_REST,482_URL./,482_URL./noticias/ => LiteSpeed\LSC->send_headers()@620 => WP_Hook->apply_filters(,ARRAY)@324 => WP_Hook->do_action(ARRAY)@348 =>
/home/admin/domains/**masked**/public_html/wp-includes/load.php@517

Let me know if I’m reading it correctly:

A POST request to /wp-json/ from the post.php admin page, issued a Cache Purge. But what was purged? I assume this was triggered when saving post, correct?

Anyway, I want full control of cache purging, because this is a site with big audience and multiple authors updating a lot of stuff every minute, and I’m seeing full cache being purged at least twice per hour.

Thanks for the feedback. I’ll play with the Brotli and Gzip for both Cloudflare and local to see if the issue still happens.

@supervisior could you describe more about this encoding error when Cloudflare have a version of the cache and Litespeed have another? Maybe we can find something to mitigate it. Currently we use Cloudflare Cache because it’s simply overpowered and drastically reduces our server load. We’ve built internal API integration with Cloudflare to clear content whenever we clear LiteSpeed cache, even in special cases.

What should I look on the logs, doesn’t seems to contains anything relevant? There are entries about ESI:

[ESI] 💕 [BLock_ID] esi [wrapper] esi-shortcode [Control] public,no-vary

I see no errors when the ESI is not executed though

Ok, found the bug.

I’ve checked if the footer was triggering, and it was as intended, then I saw that on a related function (send_headers_force)

* For compatibility with those plugins have ‘Bad’ logic that forced all buffer output even it is NOT their buffer

I had this piece of code on the wp_footer action:

ob_start();
custom_post_ad($target);
$ads = ob_get_clean();
ob_end_flush();

After removing it, it worked. Do you know why this happened? Is that way of getting the buffer wrong? Or the plugin isn’t compatible with buffer retrieval at the footer action?