adjust handle count per host?

I've just recently installed Event Sentry and I'm wondering if there is a way to adjust the performance monitoring on handle count on a per-machine (or group) basis. I've got a couple of Microsoft SQL servers that always exceed the handle thresholds. I'd rather not raise that threshold on the rest of my servers if I can help it.

Is there an easy way I can tweak that threshold just for those servers (or a group)?

Comments

  • It's not currently possible to customize the thresholds for performance alerts, although this is definitely something we plan on adding in the near future.

    Otherwise you'd have to create a separate package with identical settings with the exception of the alert threshold. This will ensure that the correct thresholds are being used for both alerts and the dashboard(s) in the web reports.

    If you are mostly concerned about suppressing alerts, then it is possible to do that utilizing insertion strings and the numerical comparison feature in filters.

    You would basically create a package & exclude filter which you would only assign to the SQL Servers. The exclude filter would look for the insertion which represents the current performance value of the counter, and only match if the value exceeded a higher threshold.

    Performance alerts for counters with instances (such as processes) are usually logged as event id 12105, with insertion string number 4 (%4) representing the current average value.

    Since you want to exclude alerts coming from the SQL servers with high handle values, you would pick a threshold higher than the one configured in the system health package.

    We plan creating a tutorial illustrating this, but a couple of our screencasts http://www.eventsentry.com/support/screencasts (EventSentry's Application Scheduler Part 1 & 2) also illustrate this functionality based on a different scenario.

    I hope this helps, let us know if you need more information on how to set this up.
Sign In or Register to comment.