Custom Query (2297 matches)

If a post body contains \x00 (or maybe other control characters) the variable $request_body is incomplete.

Two examples:

• a post request with the content BEFORE_NULL\x00AFTER_NULL - the variable $request_body contains only BEFORE_NULL.
• a post request with the content SOME_PREFIX\x00... some more content \x01 ...MESSAGE CONTINUES HERE - the variable $request_body contains only SOME PREFIX MESSAGE CONTINUES HERE

I'm not sure of the exact mechanics of what's happening here, but I've created a minimal, reproducible example here: ​https://github.com/AD7six/nginx-partial-request-body-repro based on the current nginx:latest docker image.

Am happy to provide more information if necessary or to chat about this via the community nginx slack ​https://nginxcommunity.slack.com/archives/C02T18NMZU4/p1665412094085809.

Если в конфигурации есть несколько секций server, и хотя бы в одной из них в listen указано ssl http2, то и для других секций с SSL используется HTTP/2.0 для клиентов, даже если в них не указано http2.

А по умолчанию такое поведение не ожидается.

We use (and love!) nginx to do rate limiting on calls to some of our servers. We'd love the ability to configure the error that gets returned when customers are rate limited (limit_req module). This message should be configured independently from other server errors.

There is a new draft spec RFC6585 ​http://tools.ietf.org/html/rfc6585 that specifies http status code 429 Too Many Requests which is perfect. Since this is draft many not everybody wants this as the default, but at least being able to configure it would be great.

My understanding is that using error_page 503 = 429 /too-many-requests.html would change the behavior for all 503 responses, which isn't correct. If there is a problem with my server, I don't want to tell customers they are being rate limited.

FYI, I have an open question about this on ServerFault: ​http://serverfault.com/questions/483798/how-do-i-configure-nginx-to-return-429-http-code-when-rate-limiting