HTTP 500 Internal Server Error: find the cause in the logs

What the message looks like

PHP Fatal error:  Allowed memory size of 134217728 bytes exhausted (tried to allocate 20480 bytes) in /var/www/site/vendor/autoload_real.php on line 41

Common causes

• PHP fatal error, e.g. “Allowed memory size … exhausted” or an undefined function after an update.
• Missing PHP extension after switching PHP versions.
• The application cannot write to cache, log or upload directories.
• Rewrite loop in nginx (“rewrite or internal redirection cycle”).
• Database unreachable or credentials changed.

Check

1. Show recent errors in the nginx error.log
   tail -n 50 /var/log/nginx/error.logCopy
2. Show PHP-FPM messages from the last minutes
   journalctl -u 'php*-fpm' --since '15 min ago' --no-pager | tail -n 30Copy
3. Search the PHP-FPM log for fatal errors
   grep -i 'fatal' /var/log/php*-fpm.log | tail -n 10Copy
4. Fetch the status code reproducibly
   curl -sS -o /dev/null -w '%{http_code}\n' https://example.com/Copy

Fix

• Read the first error at the time of the request – it names file and line.
• For memory errors raise memory_limit in the pool or php.ini deliberately, not blanket -1.
• Install missing modules (e.g. php8.3-intl) and restart PHP-FPM.
• Give the PHP-FPM user write access to cache and log directories.
• After a failed update roll back to the last working version and analyse calmly.

FAQ

Should I turn on display_errors?

Not in production. Errors in the browser leak paths and credentials. Keep log_errors = On and read the log.

Why is there nothing in the nginx log?

Then PHP handled the error itself and only returned 500. Check the application log or the PHP-FPM log with catch_workers_output = yes.