November 5, 2012 at 5:00 am
Hi,
Is there a way to detect from within SQL (i.e. not using perf mon counters) that my system is suffering from memory pressure?
For example for CPU I have used sys.dm_os_ring_buffers DMV (particularly buffer type: RING_BUFFER_SCHEDULER_MONITOR) to detect CPU over-utilization if the ring_buffer sampling rate showed 3 consecutive entries above 80% utilization. Is there some way to do the same with memory?
I have tried:
> Using sys.dm_os_ring_buffers DMV with RING_BUFFER_RESOURCE_MONITOR: Doesn't seem to be as sensitive (or accurate?) as I would like it to be, I know the box is suffering from memory issues but this DMV is not reporting anything.
> Using the dm_os_memory_clerks DMV (as highlighted in this article): I think this particular DMV has changed since SQL 2005, and am not sure how to map the new column to the old columns mentioned in the article.
Any suggestions/guidance on what I should use?, or should I abandon this concept and start looking at perf mon counters instead?
Thanks!
-----------------
... Then again, I could be totally wrong! Check the answer.
Check out posting guidelines here for faster more precise answers[/url].
I believe in Codd
... and Thinknook is my Chamber of Understanding
Viewing 0 posts
You must be logged in to reply to this topic. Login to reply