504 Gateway Timeout

An HTTP 504 Gateway Timeout error means that a server acting as a gateway or proxy did not receive a timely response from the upstream server it needed to access to complete the request. This is distinct from a 502 in that the upstream server did not respond at all within the allowed time, rather than sending a bad response. End users experience a long wait followed by an error, while developers should investigate latency in the backend.

Common causes

• Slow database queries or unoptimized API calls causing the upstream server to exceed the proxy timeout window
• Network latency or routing issues between the gateway and the upstream server
• The upstream server is under heavy load and cannot process requests within the configured timeout period
• Proxy or load balancer timeout settings are too aggressive for the expected response time of the backend
• DNS resolution delays or intermittent DNS failures when the gateway tries to reach the upstream server

How to fix it

• Identify and optimize slow backend operations such as database queries, external API calls, or heavy computations
• Increase gateway timeout settings (e.g., proxy_read_timeout, proxy_connect_timeout) to allow more time for slow endpoints
• Add caching layers (CDN, Redis, Varnish) to reduce the number of requests that reach the backend
• Implement asynchronous processing for long-running tasks and return a 202 Accepted with a polling endpoint instead
• Check network connectivity and latency between the proxy and upstream servers using tools like traceroute or mtr

Related HTTP errors