It seems that OAM is parsing the wrong field somewhere when gathering CPU usage info. It affects the getsystemcpu command, and others that return overall CPU usage. Those commands always return 0% cpu usage even when I load them down with dummy processes (ex, some bash that does useless stuff in an infinite loop).
I suspect this is consistent across environments, but fwiw, the environment I found this in is a DBS-like installation. Centos 7 VMs. Originally installed 1.1.6 non-distributed and non-root. Then upgraded to 1.2.1 with non-root/non-sudo/non-distributed. 2 nodes, 1 um, and 1 pm.