502 Bad Gateway means a server acting as a go-between, usually Nginx, a load balancer or a CDN, passed your request to the program behind it and got an invalid answer or none at all. The go-between is fine; the thing behind it is not.
If you are visiting the site
The problem is on the website’s side. Reload after a minute, since a 502 is often brief while a service restarts. If it persists, try again later, or check whether others report the site is down.
If you run the site
Work from the go-between backwards. Most 502s have one of these causes.
1. Is the backend running?
For PHP sites the backend is PHP-FPM; for apps it is the app’s own process.
systemctl status php8.3-fpm
systemctl status php-fpmThe first name is used on Ubuntu and Debian (your version number may differ), the second on RHEL-based systems. If it is stopped or failing, start it and read why it stopped:
sudo systemctl restart php8.3-fpm
sudo journalctl -u php8.3-fpm -n 502. Read the web server’s error log
The log says exactly what went wrong:
sudo tail -n 50 /var/log/nginx/error.log| Log says | Meaning |
|---|---|
connect() to unix:/run/php/... failed (2: No such file or directory) | The socket path in Nginx does not match PHP-FPM’s |
connect() failed (111: Connection refused) | Nothing listening on that port; app is down |
connect() ... failed (13: Permission denied) | Socket permissions, or SELinux blocking the connection |
upstream prematurely closed connection | The backend crashed while handling the request |
3. Check the socket or port matches
Compare fastcgi_pass (or proxy_pass) in your Nginx site with the listen line in the PHP-FPM pool file, usually under /etc/php/8.3/fpm/pool.d/ or /etc/php-fpm.d/. They must point at the same socket or port.
4. Check memory
When a server runs out of memory, the kernel kills processes, and PHP or the app is often first.
free -h
sudo dmesg -T | grep -i 'killed process'If you see killed processes, add memory, add swap, or lower PHP-FPM’s pm.max_children.
5. SELinux on RHEL-based systems
On AlmaLinux, Rocky and RHEL, Nginx may not connect to a network port until you allow it:
sudo setsebool -P httpd_can_network_connect 1Behind a CDN or load balancer
If the 502 page is branded by your CDN, the CDN could not reach your server at all. Check the server is up, that its firewall allows the CDN, and that the CDN points at the right IP.
Related
- 504 Gateway Timeout is the slow cousin: the backend answered too late.
- What a reverse proxy is.
Something out of date? Software changes. If a step no longer works, tell us and we will check it and update the page.
