There are multiple threads connected to a given DB2 subsystem. When trying to bring down that DB2, these threads seem to hold onto the subsystem keeping this DB2 from coming down quickly. Should the datacollector be brought down for the member before bringing down that DB2?
OMEGAMON shows the the same threads. When the normal -STOP DB2 command is issued, OMEGAMON lets go and DB2 comes down.
Release : 19.0
Component : CA Insight Database Performance Monitor for DB for z/OS
The recommendation is to bring down Datacollector first. Here is the link from the doc about stopping the datacollector: