justind wrote:
I just got started with zenoss today. Grabbed the vmware image for 2.2.4. Most
of the basics are pretty straightforward, at least for data that goes into rrd
graphs.
Great! Glad you were able to get going so quickly.
I'm looking for a shove in the general direction I need to go, because it's
just not jumping out at me in the docs.
I have a few string/ip address values stored on some endpoints that can be
retrieved via snmp.
Things like software version, call agent ip address, etc. I'd like to store
these values and have them displayed on (preferably) the Status screen of the
device.
It would also be handy to display the current status of a few other values that
I can get via SNMP, but no need to graph them.
For example:
I have devices that have and integer value for PowerStatus.
0 = Normal
1 = Low Battery
2 = AC Power Fail
I can raise events with thresholds, that's very straightforward. However, when
troubleshooting a device it would be nice to display this values status right on the
status screen. This particular value we don't want to have raise an event, or even change
the "severity" of the device. We just want the information readily available.
I also have a couple of questions regarding severity:
1) Is it possible to define more custom levels of severity?
2) Is it possible to redefine the severity colors displayed on either the
zenoss map or more importantly, the google maps display (I can't tell red from
green on google maps)
Zenoss has a two-phase data-collection engine. The first stage is
called modeling and the second is called performance collection. Data
that rarely changes (such as the number of interfaces in the device,
etc) is collected during the modeling phase (run every 12 hours) and is
stored in zodb. Rapidly changing data is collected in the performance
phase (run every 5 minutes) and is stored in rrd. Modeling data is
informational, where performance data is used for generating graphs and
events.
You can update the Status page during the modeling process, but my guess
is that this does not run frequently enough for your needs. However, a
performance collection cannot adjust the model by itself. The
work-around would be to use the performance data to generate an event.
Then you have an event command which runs when that event is created
which would update the model (using something like zendmd, or perhaps
REST) and then move the event to history.
Hope this helps.
Nathaniel
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
