504 Gateway Timeout

The short version: Reload after a minute before anything else: a 504 is almost always temporary and lives on the website's side, not your device. It means a gateway or proxy gave up waiting on a slow or overloaded upstream server. Run the website yourself? Speed up that back end or raise the gateway timeout instead of papering over the delay.

Key takeaways

• A gateway or proxy reached an upstream server but the answer never arrived in time, so it stopped waiting. RFC 9110 puts it plainly: the server did not get a timely response from an inbound server it needed to finish the request.
• This is the website's problem, not yours. As a visitor you have nothing to fix. Wait a minute and reload.
• 502 and 504 split cleanly. A 502 means the upstream replied with garbage. A 504 means it never replied in time. Compare those with a 503 (the server refusing on purpose) and a 500 (the application crashing).
• Behind Cloudflare, an origin that times out returns 524, which fires at 120s, not 504. So a plain 504 coming through Cloudflare usually traces to a second proxy or the origin's own gateway.
• For owners the order matters: fix the slow upstream first, then reach for a longer gateway timeout (nginx proxy_read_timeout, Apache ProxyTimeout) as a stopgap. Longer timeouts never cure slow code.

Error type

HTTP 5xx server error

Whose side

Almost always the website

Fix difficulty

Moderate (website owner)

Common cause

Upstream server too slow or overloaded

What does 504 Gateway Timeout mean?

A 504 Gateway Timeout is an HTTP status code that shows up when one server, acting as a gateway or proxy, does not get a response in time from another server it depends on. The server you reached is online; it simply gave up waiting on a back-end server to answer.

Because the timeout happens between two servers, a 504 is almost always a problem with the website's infrastructure, not your browser, device, or internet connection. As a visitor, there is little to do beyond waiting and reloading.

504

HTTP status code

~60s

Default Nginx proxy timeout

120s

Cloudflare origin timeout (Error 524)

How the 504 Gateway Timeout error appears

The wording changes depending on your browser, device, or server. Here is how this error commonly shows up:

example.com

504

Gateway Timeout

The server didn't respond in time.

• 504 Gateway Timeout
• HTTP Error 504
• Gateway Timeout (504)
• nginx: 504 Gateway Time-out
• Cloudflare: 504 Gateway time-out (Error 504)
• This page isn't working / took too long to respond

504 vs 502, 503 and 500

All four are 5xx server-side errors, but each points to a different failure between the gateway, the origin, and the application.

Gateway and upstream timeout directives

A 504 is almost always one of these timers expiring before the upstream answered, so this is where owners look first.

What causes 504 Gateway Timeout?

• The origin (upstream) server is overloaded and answering too slowly, which often happens during a traffic spike.
• A slow database query or a long-running request runs past the time the gateway is willing to wait.
• The origin server has crashed or gone offline, so the gateway gets nothing back at all.
• Reverse proxy or load balancer timeouts (Nginx proxy_read_timeout, Apache ProxyTimeout, HAProxy, and so on) are set tighter than the request actually needs.
• A CDN or proxy in front of the website, such as Cloudflare or Fastly, cannot reach the origin inside its timeout window.
• Network or firewall trouble between the gateway and the upstream server delays or drops the connection.

How to find the cause fast

1. Reload once after a minute, then open the website on another network or through Is Website Down? to confirm it is failing for everyone, not just you.
2. Owners: read the proxy and origin logs (for example tail -f /var/log/nginx/error.log) and look for upstream timeout entries. That tells you whether the proxy or the back end is the one timing out.
3. Time the request with curl -w "%{time_starttransfer}" -o /dev/null -s https://yourdomain.com. A slow upstream and an unreachable one look very different here.
4. Got a CDN or proxy out front? Hit the origin directly (or pause Cloudflare) to learn whether the timeout sits at the edge or the origin.

Terminal

curl -sI https://example.comHTTP/1.1 504 Gateway Time-outServer: nginx# a healthy origin replies in time: HTTP/1.1 200 OK

How 504 Gateway Timeout looks from the outside

Seen from outside, a 504 is an answered request, not a dead one. An external monitor logs a clean HTTP 504 with a fairly normal response time, and that alone tells you the edge is up while an upstream server is the culprit, well before you open a log. These failures tend to be spiky and short-lived in our experience, so a single failed check is usually just noise. The same 504 repeating across several locations for a few minutes is the signal worth acting on.

To confirm the exact code a URL returns, or to re-test several at once after a fix, run them through the free bulk URL status checker.

How to fix 504 Gateway Timeout

If you are a visitor

1. Give it a minute, then reload the page. A 504 is often temporary and clears once the server catches up.
2. Try the website in a different browser or a private window to rule out a cached error page.
3. See whether it is down for everyone with Is Website Down? If it is down for all, the issue is on their end and waiting is your only option.
4. Only look at your own connection if other websites are failing too. A 504 almost never points back at your network.

If you run the website

1. Check the origin server's health and logs for crashes, out-of-memory events, or processes stuck at 100% CPU.
2. Hunt for slow database queries or long-running application requests. A single slow query can hold a worker long enough to trip the timeout.
3. Confirm your reverse proxy or load balancer can actually reach the upstream app server, and that the app is listening on the port it should be.
4. When the work legitimately runs long, raise the gateway timeout (Nginx proxy_read_timeout and proxy_send_timeout, Apache ProxyTimeout, HAProxy timeout server). This is a stopgap, not a cure for slow code.
5. On Cloudflare or another proxy, verify it can reach your origin. Keep in mind Cloudflare's own origin timeout is Error 524; a 504 through Cloudflare usually means a gateway between it and your origin, or the origin itself, handed back a 504.
6. When the timeouts track your traffic peaks, add capacity or scale out so the origin can keep pace.

Still not fixed? Next steps

• When a longer gateway timeout only pushes the 504 back a few seconds, go after the upstream itself. One slow database query or external API call can hold a worker hostage past every timer you set.
• Hit the origin directly. If it answers fine but visitors still see 504, the fault lives in the proxy, load balancer, or CDN config between you and them, not in the application.
• Do the 504s track your traffic peaks? Then the upstream is saturated. Add workers or scale out instead of just stretching timeouts.
• On a managed stack where none of this applies, hand your host the affected URL plus exact timestamps. That lets them inspect the gateway and origin from their side.

Code & configuration

Copy-paste starting points. Replace example.com and the paths with your own, and test changes on staging before production.

Raise the gateway timeout (Nginx)

location / {
    proxy_read_timeout    120s;
    proxy_connect_timeout 120s;
    proxy_send_timeout    120s;
    fastcgi_read_timeout  120s;
}

Apache and PHP equivalents

# Apache (httpd.conf)
Timeout 120
ProxyTimeout 120

# PHP (php.ini): for long-running scripts
max_execution_time = 120

How to prevent 504 Gateway Timeout

A 504 tends to come and go inside a few minutes, which is exactly why it slips by until a customer flags it. Pulsetic checks your website from multiple locations as often as every 30 seconds and alerts you by email, SMS, voice call, Slack, Discord, and more the instant it starts returning 504s or stops responding. It will not name the slow query for you, but it tells you fast that something is wrong, and when it recovers, so you can move before your visitors do.

Learn how Pulsetic's uptime monitoring detects this from the outside, across 15+ locations.