SQL Server WMI Event Alert and BROKER_TASK_STOP waits

Question

In SQL Server 2008 I created a Sql Server Agent Alert last week to monitor deadlock events. The Alert calls a job with no schedule sending the xml deadlock graph token obtained through WMI SELECT * FROM DEADLOCK_GRAPH. The job calls a SP that stores the graph and an email is sent.

I have been monitoring the Sql Server Waits and have found that after I implemented the alert described above, the wait categories for BROKER_TASK_STOP and SQLTRACE_WAIT_ENTRIES have sky rocked. Those values went from not even being on the radar to contributing 42.59% and 5.11% of total waits respectively.

Should I be concerned? I read somewhere that high BROKER_TASK_STOP's may not be of real concern.

Answer 1

These two wait types are excluded from Paul Randal's wait stats analysis query, and he is far more of an authority than me.

I would not be concerned about these, and exclude them from your analysis.

Answer 2

These are both 'friendly' waits and nothingbtonworry about from a performance perspective. They mean that threads are sitting there waiting for work to be assigned. You should filterboutw these and other friendly waits from your wait stats queries. Many have blogged about which waits to exclude, including SQLRockStart, here - http://thomaslarock.com/2012/05/are-you-using-the-right-sql-server-performance-metrics/