After updating to Nginx 1.30.0 I started noticing random, hard‑to‑debug issues. Using PHP (FPM) upstream via sockets (no keepalive defined), with fastcgi_keep_conn enabled, PHP workers would hang unpredictably, causing noticeable lag. Disabling fastcgi_keep_conn immediately resolved the problem. Since 1.30.0 introduced major upstream changes, is this behavior expected?
What version were you using previously?
Are you getting any errors in the nginx or php-fpm logs?
I always keep Nginx updated to the latest stable release. Before 1.30.0 I was running 1.28.3. In the Nginx error logs I consistently saw 504 timeout errors. To investigate further I enabled the PHP-FPM slow log with a 5‑second threshold, but the traces were very random and dispersed, without any conclusive pattern.