(Story 1) At a government site, EPV highlighted an IMS transaction that had jumped to a much bigger number of executions, causing a usage of 150 MIPS in the peak hours. After a quick check with the people supporting the application, Performance Analysts discovered the “wake up” parameter had been mistakenly set to zero causing the transaction to continuously run, searching for work to do. By setting the correct parameter the MIPS used dropped to only 10.

(Story 2) EPV highlighted a statistically abnormal response time for a business critical CICS transaction. A further analysis showed the increase was in the time spent in DB2 doing a table space scan. Looking at DB2 statistics the DBAs realized that the day before a runstat had been performed when the table was empty.