collector/proxy implementation.

[Giray Devlet]

Hi ...

  I haven't been able to find anything like a seperate
nagios collector / proxy.

  The idea is to have a seperate collector that gathers
information from nrpe's.  Then nagios gets its information
from the collector (via a seperate plugin, similar to
check_nrpe).

  The need for something like this comes from the need
for scalability (to have seperate collectors for seperate
environments).  Or to use it as a gateway between network
borders. Is there anything similare being done?

  My idea is to borrow heavily from nrpe-2.0 and do the 
following:

   - 'check_collector' is a plugin for nagios, it
       communicates with a given collector.  Has parameters
       such as collector-IP, host-IP, check-command.
   - 'collector'  is a daemon that can work in two modes
          I) transparent
         II) non-transparent

      * In transparent mode it will receive requests from
         a specified nagios host, and blindly forward the
         request to the specified  host with the given command.

         Received requests will be cached, and continuously
         (according to a specified interval) checked.  Results
         will be cached at the collector.  When nagios asks
         for a status, cached information will be provided.
   
         If for a certain period new requests do not arrive, 
         the check for that service will not be repeated. 

         Advantage: No need for extra configuration on the 
                        collector.  (only specification of authorized
                        nagios host)
         Disadvantage:  
                        - nagios host IP can be spoofed and random
                           checks can be executed. DOS attacks!
                        - initial check will probably result in timeout.

      * In the Non-transparent mode extra configuration 
         required. host:command pairs need to be entered 
         to specify which checks are allowed.

         checks will continue independent from requests generated
         by the nagios host.  When a request is made, cached 
         information will be provided.

         Advantage: In case of a DOS attack, monitored systems
                        will not be directly bothered.
         Disadvantage:  One more system will have to be configured.

Any ideas, comments and pointers welcome...

gd

        


  



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf