[Nagiosplug-help] RE: check_http, but click-through

[Ralph.Grothe at itdz-berlin.de]

Hi,
 
since I have subscribed to this mailing list only recently (I'm
still new to Nagios) I have been receiving quite a few mails
regarding the plug-in check_http.

As I am still in the phase of exploring the plug-ins I also
stumbled over check_http because it seemed the right plug-in for
our webservers.
Like a well behaved Nagios plug-in it nicely displays the options
and arguments it could care for when invoked with --help switch.

Unfortunately I need to check some webservers that serve to the
WWW from there which requires me to go over our WWW proxy.
I haven't seen a switch from check_http's help screen that would
provide for a proxy server.
That's why I am forced to write my own plug-in.
However, this shouldn't be too dificult since Perl's LWP module
gives you an easy interface with full command over e.g.
formatting of HTTP headers for such things as HTTP proxies.

On the other hand I cannot beleive that my requirements being
overly off the road,
and I'm sure someone must already have written a Nagios plug-in
that also honours HTTP proxies.
Or I simply have missed the (hidden) features of check_http
(haven't looked at the source).

Can someone give me a hint how to monitor webservers over
proxies?
How should the dependencies be treated?
Does it make sense at all because a not OK return code could
simply mean that the proxy ceased serving
while the targeted webserver might still be kicking?
If one instead should use an internal route to DMZ webservers how
would one overcome limitted open ports
(e.g. would SSH tunneling be an option, or again too many
dependencies)?

Rgds.
Ralph