Home > Self Tuning Thresholds > Self Tuning Thresholds

Self Tuning Thresholds

Sadly they haven’t quite lived up to the hype and expectation. Hopefully work will continue on them as the underlying principal is undoubtedly very useful. It is just the actual implementation in the current version hasn’t hit the mark:

http://blogs.technet.com/b/kevinholman/archive/2008/03/19/self-tuning-thresholds-love-and-hate.aspx

http://social.technet.microsoft.com/forums/en-us/operationsmanagermgmtpacks/thread/3E867995-8ABF-4084-A058-192B795212F8

http://social.technet.microsoft.com/Forums/en/operationsmanagergeneral/thread/688ee47f-4882-4f11-894f-42d935501290

Just in case the links get groomed over time, here is a summary of some of the limitations as stated by Kevin Holman:
1) “If the healthservice is ever restarted on the agent, the learning process must begin again before you will get an alert.”

2) “These dont always work well for all counters especially counters that vary widely or are ever at zero. You will find the STT to create many alerts when there is not an actionable situation. Also – the STT will not alert until the learning period is set. The learning period is reset on each Health service restart…. so if your learning period is 1 week, and you reboot your server monthly for patching – you only get monitoring 3 weeks out of the month.

In general, as Ake points out in one of the threads – “An alternative could be a consecutive samples over threshold monitor.”

Advertisements
  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: