NGINX will not reload, or it answers 502 for a site that used to work.
nginx -t before a reload. A 502 is usually the upstream, not NGINX itself. The VPS stays in your provider account.
Config, certificate, or the socket behind it
A vhost edited by hand, a duplicate server name, a certificate path the panel moved. The error is in the test. We do not reload a config that fails it.
PHP-FPM down, the socket path changed, or the upstream timed out. NGINX is doing what it was told. The fix is the pool, unless the proxy block is wrong.
Hestia, DirectAdmin and Vesta regenerate vhosts. An edit in the generated file disappears. We change the template, then rebuild one site.
A slow WordPress page is not an NGINX bug until the access log says so
Static files and a reverse proxy are in scope. Theme code is not. Cache for anonymous pages can sit in NGINX. Cart and admin stay uncached. Apache in front of the same site is a different page.
€55 an hour. A 502 with the site down is emergency if someone is free.
Minimum 1 hour. Emergency €95, subject to availability. A new VPS with NGINX, PHP-FPM and a certificate starts at €149.
NGINX failed a reload, or the site is 502?
Tell us the panel and the status code. No root password in the first message.