Tweak proxy_next_upstream based on HTTP method

Server NGINX
1

Hi all,

I was wondering if there is a way to have different proxy_* rules
depending
on the HTTP method? My use case is that I want to be a little more
conservative about what requests I retry for POST requests, as they have
an
undesired impact if tried twice after a “false” timeout.

e.g., for GET, I may want to try another when I timeout to the back-end
server after 5s, whereas for a POST request, I may want to simply fail
the
POST request after 15s.

I can work around this by doing specific location blocks for each URL,
but
having separate defaults based on HTTP method should be easier to
maintain.

Thanks,
Branden

2

Hello!

On Fri, Jul 05, 2013 at 08:32:38AM -0400, Branden V. wrote:

I can work around this by doing specific location blocks for each URL, but
having separate defaults based on HTTP method should be easier to maintain.

There is the limit_except directive which might be helpful, see
Module ngx_http_core_module.

–
Maxim D.
http://nginx.org/en/donation.html

3

Thanks for the quick reply Maxim. That looks interesting, though in
particular proxy_next_upstream and proxy_read_timeout don’t report to be
valid in that context. I’ll give it a try, perhaps it’s just an error in
the docs.

4

No it doesn’t look like limit_except will allow me to change behaviour
of
proxy timeouts or next_upstream based on the HTTP method. Are there any
recommendations on how this could be done aside from my workaround?