The way Zenoss does service monitoring built in won't work for you as it
would expect a particular service to be on a particular computer and
error if it was not.
I think you'll want to use a command datasource (well, that's how I
would do it anyway) and WMIC to poll the WMI list of services, then you
can search through it... You can use any scripting/programming language
your host OS supports/has installed, so you could use python, shell
scripting, awk, c++, pretty much whatever. Then you'll want to format
the output for Zenoss, which is basically the Nagios plugin output
format. I recall that if you're not graphing it, you can actually just
use output sort of like check_memcached does:
MEMCACHED OK - OK, Size checked; OK - at www.example.com:11211 |
size=0.00[%];90;95
So for instance it would alert if it was not OK. The trick of course is
passing the service(s) not running back, I'd look at the admin guide for
some of this... I haven't tried what you're trying to do, but I think
inside the perf template you can both define alert criticality, as well
as use various variables to create the event subject/details... which
then can be put in the e-mail alert. Hopefully someone who has done this
sort of custom command can chime in with details (and if I'm going down
the primrose path here - there might be a better way to do this than a
command datasource).
--
James Pulver
Information Technology Area Supervisor
LEPP Computer Group
Cornell University
eshamow wrote:
Something of an odd monitoring situation --
I've inherited a cluster of Win servers running services which don't have any
pattern to their name. Unfortunately the only way I can identify these
services is by the text in the Description field.
The current monitoring environment is WhatsUp Gold; to test these services
there is an ActiveScript which collects the full list of services from WMI,
searches for the subset which contain our key word in the Description field,
and returns errors if any of the services are down (along with the name of said
services).
The challenge is that the services move from server to server, so I can't just
identify the subset once -- it must be done essentially on every poll.
I'm trying to migrate this to Zenoss. How would I replicate this logic (or is
there a better way to do it) within Zenoss?
Thanks to any who can assist.
-------------------- m2f --------------------
Read this topic online here:
http://community.zenoss.com/forums/viewtopic.php?p=22990#22990
-------------------- m2f --------------------
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
