[collectd] Tomcat plugin

[Florian Forster]

Hi Justo,

I've fixed the problems you've found, I've fixed them.

On Thu, Dec 20, 2007 at 01:38:59PM +0100, Justo Alonso wrote:
> About it ... in the first message I send you a collectd servlet. If
> you don't want stats about garbage collector then you can use the
> manager/status servlet that is bundled with tomcat.

Sorry if I ask absolutely newbie questions, but I am not familiar with
Tomcat. I know it's a kind of webserver for Java applications, and
that's all ;) Could you please provide a short but informative
description that I could put in the manpage? It should contain:
- A note that the standard `manager/status' servlet can be used and a
  short configuration snippet or a link where such information can be
  found in the Tomcat documentation.
- A note what the adapted servlet does different.
- Build instructions for the servlet.
- A sample tomcat configuration for the servlet.

Also, what should I put into contrib/? I do _not_ want to distribute any
binaries, to I assume the two *.java files? Do I really need all that
directory hierarchy? I don't like having an unnecessary directory depth.

> I see the web git repository and I can't find this branch to see the
> man page ... can you send me a copy for read it ?? ;-)

That's because I wrote it just now ;) Here's what I have so far:
-- 8< --
=head2 Plugin C<tomcat>

The C<tomcat plugin> reads a status page provided by the standard
C<manager/status> servlet or a custom servlet for collectd which provides more
information (see below). The page is read using C<libcurl> and parsed using
C<libxml2>, so you need these two libraries for the plugin to work.

The information collected includes information on "connectors" (number of busy
and idle threads, number of workers in various states and number of requests
and bytes) and general memory utilization. If you use the servlet that was
adapted for collectd also information about the garbage collection (number of
times the garbage collection ran and time spent doing garbage collection) is
collected.

Since you may have several Tomcat instances running on one machine, the
configuration is organized in "Instance" blocks and may look somewhat like
this:

  <Plugin tomcat>
    <Instance "customer_a">
      URL "http://localhost:8001/manager/status"
      User "customer_a"
      Password "aquaifez"
    </Instance>
    <Instance "customer_b">
      URL "http://localhost:8002/collectd/status"
      User "customer_b"
      Password "daheehae"
    </Instance>
  </Plugin>

The configuration options are:

=over 4

=item B<URL> I<Address>

Sets the address of the status page. B<Required!>

=item B<User> I<Name>

Sets the user name if required. Don't specify if no authentication is needed.

=item B<Password> I<Secret>

Sets the password if required. Don't specify if no authentication is needed.

=item B<CACert> I<File>

Sets the file which holds the certification authority's certificate. This may
be required to validate the server's certificate.

=back
-- >8 --

Regards,
-octo
-- 
Florian octo Forster
Hacker in training
GnuPG: 0x91523C3D
http://verplant.org/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://mailman.verplant.org/pipermail/collectd/attachments/20071226/8ae3a5ef/attachment.pgp