Site Health in WordPress is showing a few cURL error 28 errors when we have CiviCRM installed and activated. Deactivate CiviCRM, the problem disappears. Activate it, the problem is back.
I've seen the WordPress Site Health contains critical errors unless CiviCRM is disabled question, but my research has found no justification for the top response that the health checks are inaccurate. So I need to dig deeper to try to solve this.
During a long chat session with tech support for the hosting company (SiteGround), they could find no reason for the error on their end. Server memory is maxed out, phpinfo shows everything is current and up to date (and shows server memory as 768M). So far, all of the diagnostic things I've found online have been dead ends.
With Query Monitor installed, I can see that the offending call is Health_Check_Site_Status->get_test_rest_availability(), which uses the URL
https://[URL]/wp-json/wp/v2/types/post?context=edit
That health check works fine with CiviCRM disabled, but chokes with it activated, giving the cURL error 28.
What is CiviCRM doing that causes that call to fail?