http://www.cisco.com/warp/public/477/SNMP/collect_cpu_util_snmp.html
According to this article, cpmCPUTotal5minRev of the CISCO-PROCESS-MIB should be used for multi-CPU devices. On a dual-SUP cat6k, I find this returns four CPU elements, one CPU of Routing Processor 7, one CPU of Switching Processor 7, then two on a FlexWAN card: CPU 0 & 1 of WS-X6182-2PA. OTOH, the RP and SP CPUs on the supervisor in slot 8 are not seen at all, presumably because the sup is in standby mode. Meanwhile, the avgBusy5 value matches that of cpmCPUTotal5minRev for "Routing Processor 7" as well as of "show process cpu" on the CLI. So it appears avgBusy5 is fine for monitoring multi-CPU devices too, since people will use the CLI show command to verify when they get alerted on a high CPU util situation. Am I misunderstanding the definition of "multiple CPUs"? Is it referring to the RP and SP on a single SUP, or CPUs across multiple SUPs, or a combination of SUP- and linecard-CPUs? What exactly are the benefits of monitoring a SUP's Switching Proc CPU and/or the FlexWAN CPUs, when there's no CLI method of doublechecking the util?