Waits and Queues

Waits are the delays caused when a process or worker thread cannot immediately access the resources it requires. Waits can be caused by either excessive activity or inadequate resources. Queues store incoming requests for resources that cannot be immediately satisfied.

Waits and queues can help to identify the resources which are under pressure.

But, I’m sure you already all know this, right?

What we’d like to know is – do you use waits and queues to analyze performance? If so, what specific information about waits and queues should SQL Response be giving you?

For example:
• Longest waits in last x hours e.g. top 10 waits in last 1 hour?
• Total wait time for each of these waits?

What overall performance metric for queues would be of most use to you?

As ever, let us know your thoughts.

This entry was posted in Metrics. Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.


  1. Jonathan Allen
    Posted October 20, 2009 at 3:15 pm | Permalink

    Speaking from the point of view of someone with a lot of things on my list already I have to admit that I dont monitor waits and queues as part of my regular routine. They are something that I would look to check as I investgate the server condition if there is a performance issue and its arrived without showing up on my radar in any other way. I think it would be good to have a place to reference the details ready and waiting for me, or indeed there before hand so that I could take action before the slowdown!! If there is a trend of wait stats for the past x hours/days that I can see to put the current values in context then thats useful.

    I think I would want to see current wait tasks, longest individual waits in last x hours, longest wait type (cumulative) over same period, average wait duration by wait type.

    As with all things, a basic desription of the meaning of the counters, its meaning to the server – is it a bad sort or good? and what is a lot (eg WAITFOR 15000ms is just someone making TSQL pause for 15s so ignore it, WRITELOG 10000ms is a possible IO issue on a drive and there needs to be some further analysis), ways to confirm the issue / what to do next – IE crack open perfmon and review DISK IO; review CLR connections/usage; etc

    Just a thought, are we only talking about OS wait stats or Index wait stats or both?


  2. Mike Correa
    Posted October 20, 2009 at 6:06 pm | Permalink

    I could have prevented some server issues had I known, at a glance what my queue and wait times were. I need to see the longest wait in the last x hours along with its total wait time.

    This web site dedicated to user-feedback for improving your product is top-grade.
    I can’t wait to use SR2!!

    Thanks to everyone on the team.

  3. Priya.Sinha
    Posted October 21, 2009 at 9:52 am | Permalink

    Thanks a lot Jonathan. I was referring to both the OS wait stats as well as Index wait stats. Is one of them more important for you than the other?


  4. Priya.Sinha
    Posted October 21, 2009 at 3:52 pm | Permalink

    Thanks a lot Mike for your feedback. We are really thankful to everyone who gives us feedback. With all this help from you guys hopefully SQL Response V 2.0 will definitely be a top grade product. Do keep sending us your feedback.


Post a Comment

Required fields are marked *

Add an Image


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>