Monitoring a win2000 service

despe · Guest

Hello all,

I would like to monitor a windows 2000 service like we had in the "performance monitor" in a windows 2000 server, in order to have an historical view for this service.

How can i do that ?

Thanks a lot.

ablyler · Joined: 19 Mar 2002 Posts: 40 Location: Marysville, PA

are you running cacti on a win32 or linux box?

Guest · Posted: Tue Jun 08, 2004 8:27 am Post subject:

hello,

Cacti is running on a Linux Mandrake 9.0 box, but i would like to monitor a service running on a win2000 server.

ablyler · Joined: 19 Mar 2002 Posts: 40 Location: Marysville, PA

In the past I have installed NSClient on Windows Servers in order to monitor them from Cacti. You then can use the check_nt binary that is include in the Nagios Plugins package. You will need to create a script that calls check_nt and returns the results in a cacti friendly format.

kberrien · Joined: 01 Jun 2004 Posts: 17

Do you have any more details than that? I've some Win 2k, 2k3, Nt boxes I want to monitor, and so far, I can't get any graph to work via SNMP/Cacti.

ablyler · Joined: 19 Mar 2002 Posts: 40 Location: Marysville, PA

What services do with wish to monitor on Windows? I can try to create a How-To on this subject, but it would be usefull to actually find out what type of stats people would like to get out of Windows.

kberrien · Joined: 01 Jun 2004 Posts: 17

I can't say for the origionator of this thread (Despr?/guest), but I know I'd be interested in:

I see NSClient can do:
CPU load (single or multi-processors machines)
Memory load

(I can't get these working w/snmp/cacti).

As far as services, tracking DNS requests, traffic, file services, logged in users would all be nice.

I'm monitor from a linux box like the other poster...

despe · Guest

we have a prog that run on a windows2000 server and it uses the 'centura SQLBase'. I monitor this service trought the perfom tool but i would like to have an historical view for this service. i already use SNMP informant to check CPU and so on, but i don't know how to monitor a service.

Thanks
(excuse for my poor english

)

Stiletto · Joined: 13 Sep 2003 Posts: 12

Great idea on the Nagios check_nt check, it pulls data fine but I get the following error when I try to import the data to an RRD:

update /usr/local/cacti/rra/_test_cpu_138.rrd --template cpu N:CPU Load 3% (10 min average)
sh: syntax error at line 1: `(' unexpected

As you can see the result properly pulls the CPU load but the question is how do you get the "3" out of the result data? I tried an output string setup in Data Input Methods like this:

CPU Load <cpuload>% (10 min average)

put it is not working. I'm guessing a script needs to be written to strip it down? Any ideas?

-stil

FragKing · Joined: 01 Jul 2004 Posts: 28 Location: Zurich, Switzerland

kberrien, NSClient ist the tool to go for you, I think. It can pull any data which is present in the performance monitor and you can graph it with cacti.

Search the board for NSClient and you will find a guide how to create scripts and the whole data query / graph template thing in cacti.

I'll graph some things now with it, like number of viruses my Trend Viruswall reports through Perfmon

Quite nice.

/FragKing

kberrien · Joined: 01 Jun 2004 Posts: 17

melchandra · Posted: Fri Jul 02, 2004 4:24 pm Post subject:

no, you don't need full-blown nagios. Just put the nsclient on your windows machines. There there is a linux command line utility (look on the nsclient webstite http://nsclient.ready2run.nl/) that you can run from cacti to return variables. actually, this is exactly what nagios does, it makes a call to the command line utility.

kberrien · Joined: 01 Jun 2004 Posts: 17

Great, I'll give it a try next week, after the holiday weekend, and report back on the results!