|
JaybeeSQL
Yak Posting Veteran
77 Posts |
 Posted - 11/07/2012 : 08:32:58
|
Hi all,
We have a lightly-used OLTP system, split across 3 clusters, that is showing 6 sessions with wait times often in hundreds, or thousands of seconds.
They are not causing performance complaints but ARE causing alerts to fire in our monitoring system (Idera). I would be grateful for any ideas to begin addressing the underlying issue, as well as rationalising the alerts (ie possibly by increasing the 'critical' point where an email is sent to 10,000secs).
Many thanks,
JB |
|
|
jackv
Flowing Fount of Yak Knowledge
United Kingdom
1771 Posts |
Posted - 11/07/2012 : 09:46:33
|
1)Try extracting the queries which are running at the time. This could give you some clues. Analyse the queries execution plans and see if there is an heavy CPU based activity.
2) Using a query over sys.dm_os_wait_stats check for : SOS_SCHEDULER_YIELD, THREADPOOL, CX_PACKET , usually related the CPU
Jack Vamvas
--------------------
http://www.sqlserver-dba.com |
 |
|
Session CPU wait times
