New wlm activity threshold in DB2 helps to catch poorly performing queries

Thought I would call out a few WLM features that shipped in DB2 that might be of interest...

Since DB2 9.5, DB2 has offered the ACTIVITYTOTALTIME threshold to track the amount of "clock time" that an activity spent within DB2; this threshold starts from the time the activity reached DB2 until the time it finished executing and included any time spent waiting in WLM concurrency queues before execution actually started.

This threshold is very useful when you want to track/control the end-user experience in terms of response time from DB2 but it is less effective when you have active concurrency queuing going on and are trying to track badly performing activities. Because this threshold includes time spent queuing not just executing, you can get false positives as you try to capture examples of the target statement's execution.

The new ACTIVITYTOTALRUNTIME threshold introduced in DB2 alleviates this issue by only tracking the total "clock time" actually spent executing.

Recent Stories
An old Db2 Easter Egg: Setting the default isolation value for dynamic SQL

Interpreting total_extended_latch_wait_time

How do I refresh the libraries used by Db2 external routines without stopping Db2?