How to analyze long running reports
We have a complicated financial report that appears to read in a fairly large chunk of data at the beginning of the report and then it churns along for a long time. For example, 1 day's worth of transactions can take 5 minutes to process. When we do a longer range, it goes up accordingly.
I am monitoring the underlying PostgreSQL DB where the data resides and also the DB that YF itself is attached to and see very little activity during this time after the initial chunk of data has been inhaled.
Watching the Linux system (A VM with just YF running on it), which has 32 GB of memory, JVM, which has been given a max of 20 GB at the start and the 8 CPUs shows very little stress the whole time this is running. Looking at the various info_cache.jsp, info_threads.jsp, etc.. pages shows no increase over the course of time. I also ran the yfmonitor.jar to look at the graphs and nothing looks interesting.
What other approaches can be used to understand what the system is doing during this time. (Looked through ever log that was mentioned in various support posts and found no errors and very few entries.)
Thanks for any suggestions.