A reduction in MIPS in the DISTSERV / SYSSTAT bucket was observed. This change was introduced after implementing maintenance. Is this new PTF responsible for the change in behavior?
Release : 20.0
Component : Detector for DB2 for z/OS
LU03386 did make a change to the way that INDB2_CPU for call statement was calculated, so if you are getting statistics from that value, that might explain the reduction in CPU. If you are getting stats from a non-Detector source, then we have no definitive answer, but we can say that recent changes we made in reaction to Db2 management of ECSA may have had an effect. Detector matches its main control block(ITH$ENTY) to Db2's Execution Block(EB). Rather than clean-up our control blocks immediately when we saw that Db2 EB had been deallocated, we delayed that cleanup in case Db2 came to that EB and reused it. This prevented multiple allocation and releases of our control block. We had one customer confirm that they saw a reduction in CPU when they installed that PTF, which is included in LU03461, but we have no real empirical data to prove it.