Dec 17, 2009 2:44 PM
Group interfaces instead of devices
-
Like (0)
I'm exploring Zenoss to my needs and I'd like to know if there is a way to create groups of ports/interfaces in zenoss instead of devices, which I think is a better way to represent an application monitoring availability. For example, I have a huge device that connects several servers, each of one for a specific purpose. Considering that a port is down, it could or not belongs to a server which supports an application. Can you help me with this issue?
sorry but i don't understand why it's a problem for you how zenoss work,
actually in zenoss you can give device dependency (at l3 via cli)
Andrea,
I'm not saying that Zenoss is working in a wrong way, but just asking if there is a possibility to get one specific interface of a switch, for example, and group it as part of an application to be more precisely into our monitoring purpose. The objective of this question is to dismiss the possibility of false positives monitoring alerts. I mean, if there is a switch that connects various servers, don't you think that if an specific interface goes down, it could or not belongs to that application connection. I hope now you understand what I've been looking for.
Thanks in advance.
Flávia
This is a complex and interesting item already discussed in lot of posts. If you search in the forum for "layer3 dependencies" will find lot of informations. In my opinion one of the best post (in which I have discussed) is thread/12061?start=0&tstart=0
Basically zenoss manage layer3 dependencies automatically if it is able to read and learn routing tables of all of the devices involved in the communications. When not possible, you can write python code (as suggested by Consadori) although I think that this is not always good in fact I opened a "future request" (I'm enterprise customer) about the ability to manually create dependencies in "nagios style" that, in this case, is absolutely better than zenoss.
Unfortunately if you speak about the "port of the switch", you speak about "layer2 topology" and, in this case, zenoss doesn't handle layer2. Anyway I think that using layer3 with manual dependencies you should be able to reach your goal (at least partially). You won't say: if port3 of the switch is down then...but you should be able to say(at least) : if swith goes down...then suppress events for those servers attached to this switch.
I hope to have clarified your doubt.
bye
Marco
I believe the concept of 'systems' in zenoss was primarily created with the view to move on to such a feature where components and/or devices can be mixed up together to form a particular system (or service)
For exaple, consider a service call AAA (authentication, access and authorisation)
in this container you would need to have for eg:
- the ldap servers
- load balancers (if loadbalasing used)
- the ldap processes
- the switch interfaces connected to these servers
- the mysqld process on the mysql server witch servers these (if mysql is used)
- the storage switch interface to which these get their storage
If any of these fail, then the service AAA will be deemed to be 'unavailable'
Note that some other ports on the switches may fail, but shld not raise the attention of the 'directory services consultant'
This is really a feature which can greatly improve service management.
Follow Us On Twitter »
|
Latest from the Zenoss Blog » | Community | Products | Services Resources | Customers Partners | About Us | ||
Copyright © 2005-2011 Zenoss, Inc.
|
||||||||