Announcing the upstream status module

[Maxim Dounin]

Hello!

On Tue, Apr 13, 2010 at 11:26:04PM +0530, Arvind Jayaprakash wrote:

> On Apr 12, Maxim Dounin wrote:
> >Hello!
> >
> >On Mon, Apr 12, 2010 at 10:18:33PM +0530, Arvind Jayaprakash wrote:
> >
> >> This module provides a handler called upstream_status that can be used
> >> as follows:
> >> 
> >> location /foo {
> >> 	  upstream_status;
> >> }
> >> 
> >> 
> >> It reports all the upstream blocks configured for this server. For
> >> upstreams managed using the round robin (default upstream), it lists all
> >> the backends configured in a block and indicates the current status
> >> (up/down)
> >> 
> >> http://github.com/anomalizer/ngx_upstream_status
> >> 
> >> 
> >> Improvements are welcome.
> >
> >Most important improvement is to document that it actually shows 
> >configuration and status of upstreams in one particular worker 
> >process which status request happened to hit.
> 
> Done. I did not realize this until pointed it out but it now seems
> obvious.
> 
> >And, uhm, number of allocs you use to produce simple answer is 
> >really outstanding.  ;)
> 
> I completely hated it during the development. If there is an equivalent
> of asprinf() in the nginx API, or a better way, do let me know.

Usual aproach is to count needed buffer size, allocate buffer, and 
then print.  See stub status module for an example.

Maxim Dounin