I agree with most of your conclusions.
different jobs and roles make different perspectives.
My perspective for DB2 performance is mostly:
- a java application I’m seeing first time has slow response times or bad scalability
- the custumor was trying to find the problem for months
- he doesn’t know what’s going on
- it’s a german machine, he has never heared of DBMON, but is talking about query engines and CLOSSQLCSR and visual explain...
- he doesn’t want to buy any 3rd party products
- I have maybe 2 days time to analyze the problem and to recommend ways to solve it
- analyzing thousands of (often generated) SQL statements will take > 2 years
- the database navigator (or what it might be called today) crashs, because the automatic translation of the views it uses produced misformed SQL statments, mostly caused by label on text beeing too long after translation from english to german. loading the patch and install (hopefully only one day lost). Workaround: edit the statement by hand for every query...
- start of the DBMON from client never coming back on large systems. Workaround: STRDBMON from greenscreen and some filtering with interactive SQL before analyzing the data with client
I don’t know why it takes years and a couple of releases to consolidate two query engines to one and why the DBMON throws it’s data into tables with an undocumented (it could change with next PTF) layout, so you would need a client tool to bring it to some normalized form (just to fool the russians?). This tool is better than having nothing, but I would prefere to have the ability to do some investigation automatically by programms.