Posts in category distributed

Looking for distributed monitorers (not necessarily using NSClient++)

I am looking for people interested in distributed monitoring!

As some of you might know I am building some features do manage distributed and incoherent monitoring (ie. more then one monitoring tool) so I am interested in getting some real world insight into this area. So if you or anyone you know are interested in this area and want to come with feedback on some ideas I have please don't hesitate to contact me (michael <AT> medin <DOT> name).

Also on a bit more sad note, lack of updates on the wiki/forums is due to ion part preparing for NWC but also on me trying to complete some new cool(?) features which hopefully shall be out this week.

Michael Medin

New nightly 0.4.0 build as well some som information on distributed monitoring

First off Linux seems to be doing great most features just works and thus far I am pretty impressed with the "portability". I have written a page on how to build on Linux if you want to try it out for your self build/04x/linux.

The new nightly has a lot of new features as well an updated installer with I hope works better. If you have upgraded previously from 0.3.9 I would recommend downgrading/upgrading again as you will most likely have a broken setup. (or you can just upgrade again removing any boot.ini / nsclient.ini fille you may have).

Main highlight is the PythonModule? which now supports many advanced features of NSClient++ (see the included script) as well as some new remote WMI stuff. As I have said before (I hope) one of the main new features of 0.4.x is the distribution so I have started to implement that now and one step there is the new remote WMI checking thingy.

The new remote checking features works like this. First you configure a remote host: 

[/settings/targets]
my_test_xp_vm=

[/settings/targets/my_test_xp_vm]
hostname=192.168.0.123
username=\\foobar
password=foobar
protocol=none

Then the idea is that you can run checks on "this host" which will transparently be transported to the other host. Now since NRPE does not support this (natively) we have to fake this using an argument like so: 

CheckWMI target=my_test_xp MaxCrit=3 MinWarn=1 "Query:load=Select * from win32_Processor"

But the idea is that in the future you can just do: 

nscp_client --target my_test_xp --command CheckWMI  ...

And this will happen magically using the internal routing (regardless of how nsclient++ needs to do it).

For instance the idea is that the following should be "magically": 

<master node> -{NRPE}-> <main emea node> -{SSH}-> <nordics node> -{NSCP}-> <local agent> -{WMI}-> <target>

But this is pretty far down the line so dont expect anything next few days :P

Michael Medin