I’m encountering an issue with my Nginx server where HTTP header testing tools, such as http://www.webconfs.com/http-header-check.php, return a “400 Bad Request” error. However, my pages load perfectly fine in a browser and show a status code of 200 OK in Chrome’s console.
Here’s the output from the testing tool:
HTTP/1.1 400 Bad Request
Server: nginx
Date: Fri, 07 Sep 2012 09:40:09 GMT
Content-Type: text/html
Content-Length: 166
Connection: close
I have tried increasing the buffer size with the following configuration:
large_client_header_buffers 4 16k;
Despite this, the issue persists. Can someone help me understand what might be causing the “nginx 400 bad request” error and guide me on how to resolve it?
From my experience, one common reason for a “nginx 400 bad request” error is if the request headers are too large. Nginx has default limits on how much header data it can process, so if a request’s headers exceed that, the server throws a 400 error. You’ve already tried increasing the large_client_header_buffers, which is a great start. However, you might need to ensure that the increase is sufficient for the specific headers your testing tool is sending. Also, look at the client_header_buffer_size directive in your Nginx configuration, as it may also be relevant if the headers are still too large. Sometimes, increasing both buffer sizes will resolve the issue.
Good point from Devan on checking the header size. In addition to adjusting the buffer sizes, I’d recommend taking a look at your Nginx error logs for more context on the “nginx 400 bad request” error. The logs will give you deeper insight into what’s going wrong. You can find the error logs typically in /var/log/nginx/error.log, or wherever your specific Nginx configuration sets the log path. The logs should give a clearer picture of whether the issue is due to overly large headers, a syntax error in your config, or something else entirely. This additional step can often save a lot of time troubleshooting.
Both Devan and Dipen have great suggestions. Building on that, I’d also suggest testing with different tools, especially command-line utilities like curl, which gives you more control over the HTTP request and how headers are sent. Sometimes, the “nginx 400 bad request” error might be specific to the way the online tool structures its requests. If your Nginx configuration works perfectly when tested via curl or similar utilities, the issue could lie more with the testing tool itself than your server configuration. This approach helps you determine whether the issue is server-side or related to how the request is being sent.