OK. CPU utilization tracks pretty will with rows read/converted outside of the mystery range following 9:30.But now 3 additional interesting timeperiods are evident, in the red boxes. CPU utilization is much higher in the red boxes than would be predicted by rows read/converted.What’s the CPU doing at about 8:05, 8:45, 9:30? (Not forgetting the original question about 9:40 to about 10:15.)Maybe there’s other kinds of work I have included in these graphs?
It’s interesting to see how various metrics tie together (or, as the case may be, don’t).