httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@lnd.com>
Subject RE: server-status
Date Wed, 12 Apr 2000 18:30:40 GMT
Exellent suggestion, two thoughts...

1) Extend conditionally, that is add parsing for ?mod_whatnot that
will allow the given module to provide extended status.  If not 
queried, each module returns a <minimal> amount of information.
Otherwise the status page will become a nightmare.

2) Or revisit the modules you think need to extend their status.
Perhaps their own status handlers are more appropriate  What you
mention sound like MPM stats, so those are easily extended without
revisiting the API.

  This sounds an -awful- lot like our discussions of the server
config discussion.  I really think we need to look at the two from
an API perspective, instead of a single config structure that adds 
more run-time (as opposed to config-time) string testing to the 
already overburdened ap_strcmp family.


> -----Original Message-----
> From: Mike Abbott [mailto:mja@trudge.engr.sgi.com]
> Sent: Wednesday, April 12, 2000 12:57 PM
> To: new-httpd@apache.org
> Subject: server-status
> 
> 
> I think 2.0's server status feature is too specific to connections.
> There is lots of other interesting information inside Apache that I
> would like to see on the server status page.  For instance, MPMs have
> data not only on connections but also on threads (e.g., the maximum
> number of threads dexter ever needed, which can be used to re-compile
> Apache with a smaller HARD_THREAD_LIMIT to save memory).  Also the
> static content cache I wrote for 1.3 displays a bunch of 
> counters on the
> status page, but I had to patch mod_status to add the hook.
> 
> I propose making 2.0's server status report much more general 
> by adding
> a status hook that any module could register.  Then mod_status would
> simply send the HTTP response header and iterate through the 
> registered
> hook functions, allowing each one to report its own status in HTML
> directly using ap_rputs() and the like.
> 
> I haven't done it yet but I'm willing to, and to contribute a patch
> against 2.0a2 (or whatever alpha is current at the time).
> 
> Comments?
> --
> Michael J. Abbott        mja@sgi.com        
http://reality.sgi.com/mja/


Mime
View raw message