Maaz Zindani

Loading the index…

↑↓ move↵ openesc close4072 pages
Servers, nginx and Docker

nginx still returns 504 although the PHP script has a long time limit

Error504 Gateway Time-out

Short answer

set_time_limit() only controls PHP. nginx has its own clock: if PHP-FPM sends nothing for fastcgi_read_timeout (60 seconds by default), nginx closes the request and shows 504. Raise that value for the one slow location, check PHP's own limits, and move truly long work into a background job.

  • Applies to nginx 1.26, PHP-FPM 8.3, Ubuntu 24.04
  • Updated
  • By Maaz Zindani, Senior Software Engineer

Fix it fast

  1. 1Find out which cause you have

    Find which side gave up. The nginx error log says upstream timed out when nginx's own timer ended the request. Compare the time between the request and the error with the numbers below.

    sudo grep "upstream timed out" /var/log/nginx/error.log | tail -n 5
    sudo nginx -T 2>/dev/null | grep -n "fastcgi_read_timeout"
  2. 2Match what you see, then jump to its fix

    1. 1upstream timed out ... while reading response header from upstream, about 60 seconds after the request startedfastcgi_read_timeout (60 seconds) ends the wait before PHP answersFix
    2. 2The script dies with a fatal error about maximum execution time, and the PHP log shows itPHP's own time limit stops the script firstFix
    3. 3The request always dies at the same fixed time, even with set_time_limit(0)A PHP-FPM pool setting kills the workerFix
    4. 4Even a 5-minute limit is not enough, or users wait on a blank pageThe work is too long for a web requestFix

What does the error look like?

  • 504 Gateway Time-out
  • upstream timed out (110: Connection timed out) while reading response header from upstream

Cause 1 of 4

fastcgi_read_timeout (60 seconds) ends the wait before PHP answers

1Check

The directive sets how long nginx waits between two reads from the FastCGI server, and the default is 60 seconds. A script that works silently for longer is cut off, whatever set_time_limit() says.

sudo nginx -T 2>/dev/null | grep -n "fastcgi_read_timeout\|fastcgi_pass"

2Fix

Raise the timeout in the location that runs the slow script only, not for the whole site.

location ~ ^/reports/export\.php$ {
    include fastcgi_params;
    fastcgi_pass unix:/run/php/php8.3-fpm.sock;
    fastcgi_read_timeout 300s;
}

3Verify

Test the config, reload nginx and repeat the slow request.

sudo nginx -t && sudo systemctl reload nginx

Confirmed by nginx.org

Did this fix it?

Cause 2 of 4

PHP's own time limit stops the script first

1Check

The default limit is 30 seconds for web requests, or the max_execution_time value in php.ini. set_time_limit() restarts the counter from the call. On Linux, time spent in database queries and system calls does not count; on Windows it does.

php -r "echo ini_get('max_execution_time'), PHP_EOL;"

2Fix

Call set_time_limit() with a realistic value at the top of the slow script, and keep it below the nginx timeout so PHP ends the work cleanly.

<?php
set_time_limit(240); // seconds from this call; 0 removes the limit

3Verify

Run the request again and check that the maximum execution time message is gone from your PHP error log.

curl -s -o /dev/null -w '%{http_code} %{time_total}
' https://example.com/reports/export.php

Confirmed by php.net

Did this fix it?

Cause 3 of 4

A PHP-FPM pool setting kills the worker

1Check

request_terminate_timeout is a pool setting that kills a request after a time. It is off by default (0), so a value here was set by someone.

sudo grep -rn "request_terminate_timeout" /etc/php/8.3/fpm/

2Fix

Raise it above the longest allowed request, or remove the line if you want only max_execution_time to apply. Reload PHP-FPM.

; /etc/php/8.3/fpm/pool.d/www.conf
request_terminate_timeout = 300s

3Verify

Reload PHP-FPM and repeat the request.

sudo systemctl reload php8.3-fpm

Confirmed by php.net

Did this fix it?

Cause 4 of 4

The work is too long for a web request

1Check

Each running request holds one PHP-FPM worker, and the pool is capped by pm.max_children. A few long exports can fill the pool, so new requests wait.

sudo grep -rn "pm.max_children" /etc/php/8.3/fpm/pool.d/

2Fix

Start the job in the background (a queue worker or a cron-run script), return at once with a job id, and let the page poll for the result. Use the slow log to find which scripts are slow.

; /etc/php/8.3/fpm/pool.d/www.conf
request_slowlog_timeout = 10s
slowlog = /var/log/php8.3-fpm.slow.log

3Verify

Reload PHP-FPM and read the slow log after the next slow request.

sudo tail -n 40 /var/log/php8.3-fpm.slow.log

Confirmed by php.net

Did this fix it?

How do I stop it coming back?

  • Set long timeouts per location, never globally, so a hung script cannot hold every worker.
  • Keep PHP's limit a little below the nginx limit, so PHP ends the script cleanly and logs why.
  • Move exports, imports and report builds to a background job.
  • Turn on the PHP-FPM slow log in production and review it after each release.

Sources

How was this page made?

I drafted this page with AI assistance from the official documents listed under Sources. Every cause above names the document that confirms it. I have not reproduced this fix on a machine of my own yet, so read it as documented guidance, not a test report. Found a mistake? Tell me.

Which stack is behind this page?

Projects of mine built on the same technologies:

Related fixes

Servers

nginx gives up waiting for the app and returns 504

upstream timed out (110: Connection timed out) while reading response header from upstream

nginx returns 504 when the app behind it takes the request but sends nothing back within proxy_read_timeout (60 seconds by default), or when the connection cannot be opened in time.

5 causes, each with a fix

Back end

PHP stops the script after 30 seconds

Symfony\Component\ErrorHandler\Error\FatalError: Maximum execution time of 60 seconds exceeded

PHP ends a script that runs past max_execution_time (30 seconds by default for web requests).

5 causes, each with a fix

Servers

Cookies from a PHP backend vanish behind an nginx proxy

Login works when you call the backend directly but fails through nginx

nginx forwards Set-Cookie headers from the backend unchanged, so lost cookies usually come from the cookie itself: a domain or path that does not match the public URL, or Secure and SameSite flags the browser rejects.

4 causes, each with a fix

Next step

Want this handled for good?

I look after servers, apps and releases end to end. Tell me what keeps breaking.