We are doing something somewhat similar.
We need to get the actual CPU value of a device and send that to an event
command to do something with it (in a nutshell).
What I've done is create a (MinMax) threshold (severity 'Info') for the device
and set its max value to 1, meaning that an event gets created every time it
collects the CPU.
This event gets created in a custom Event Class (/Perf/CPU/Filer).
In that Event Class I have a transform:
Code:
evt.summary = evt.message.split()[6]
This is pretty much your parser .. couldn't find a cleaner way.
The 7th element in the string, in our case, is the actual CPU value.
I then set an Event Command under the Event Manager, which gets triggered any
time an event comes in for Event Class /Perf/CPU/Filer.
This command can use the ${evt/summary}, which now is the CPU value.
So you could 'echo ${evt/summary} > /tmp/size' to store it on the filesystem
for further use or whatever else ..
This is quite a workaround for something that seems should be pretty simple ..
but it does work well for us and I couldn't really think of another way (except
maybe the rrdfetch, which we're using somewhere else as well).
-------------------- m2f --------------------
Read this topic online here:
http://forums.zenoss.com/viewtopic.php?p=30570#30570
-------------------- m2f --------------------
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
