We are attempting to analyze an MQ issue that is severely impacting one of our clients. We performed a 1 minute sample and created a report in Batch. We decided that we needed to collect additional data during the issue and modified the same monitor from 60s, 6000 samples to 2m, 12000 samples against the same QMGR and we can not produce a batch report. Basically we receive SYSTEM ABEND CODE 878 REASON CODE 00000010 with 0M coded for region.
MAT will measure applications that are using MQ. So if you have CICS transactions, or batch applications that are using MQ and you measure the CICS or Batch application we will tell you thinks like queue manager, message length, the % of activity or wait that the application is incurring using or waiting MQ.
You can measure the CHIN and we should not negatively impact either the MQ subsystem or the applications that are using it. However, MAT is not designed to give overall MQ subsystems usages statistics. We will take samples for every TCB we see, we will give very basic information like TCB, some program information but not much else because we are not a subsystems monitor.
Sysview and other subsystem and systems performance monitoring products are designed to do what you are trying to do, our function is to look at a poorly performing application and tell you what it is doing with MQ. Their job is to look at a poorly performing system or subsystem to try to pinpoint the performance issue.
In the case of MQ V8 and above, IBM has significantly enhanced SMF 118 data I believe it is, to help you deal with MQ performance issues.
MAT does have enhancement requests to add CHIN and other MQ series address space support, but it isn't there yet.