Patrick,
I believe in 10.2 and above uses the aggregate. I am basing this off the following from the $SPECROOT/SS/CsVendor/Cabletron/EventDisp file:
# CPU Threshold alarms to occur based on the
# aggregate utilization of all the instances
0x00010f07 R Aprisma.EventPairTimeAttr, 0x00010f08, "0x00010f09 -:-", 0x12bce
0x00010f09 E 50 A 2, 0x00010f09
0x00010f08 E 50 C 0x00010f09
# Memory Threshold alarms to occur based on the
# aggregate utilization of all the instances
0x00010f0a R Aprisma.EventPairTimeAttr, 0x00010f0b, "0x00010f0c -:-", 0x12bcf
0x00010f0c E 50 A 2, 0x00010f0c
0x00010f0b E 50 C 0x00010f0c
For 10.1 and below, you could modify the $SPECROOT/SS/CsVendor/Cabletron/EventDisp file to do one or the other.
Reference knowledge documents How to change the SPECTRUM CPU/Memory threshold alarming to alarm on the aggregate value and not the individual instance…
Either way, it is global.
To do it for individual, I recommend to create a SpectroWatch on the attribute used for CPU utilization for that particular model. Reference knowledge document How can CA Spectrum be used to monitor and alarm on individual CPU and memory for devices?
Joe