As far as I can tell (visually, not by actually trying it), SMF89UZT is
expected to, and does, contain the delta from the previous SMF89 record;
it is not intended to be cumulative. I suggest that you take this up with
IBM service if you are truly seeing otherwise.
If you seek a change in IBM
Happy New Year to all.
I'm looking at product usage data in SMF 89 subtype 1 records, and I see that
TCB and SRB values go up and down, and look like reasonable interval values.
The SMF89UZT field for offloaded CPU time, however, only seems to go up and
makes it look suspiciously like it