SCOM Tuning Thresholds

We currently use SCOM to monitor our customers Lync/Skype for business services, we keep getting a ton of alerts normally overnight when the servers do their maintenance tasks, most of these are open alerts that then close, is there a way in SCOM to tune the threshold so it only notifies us after a certain time or on a certain number of alerts.

1 Like

Maybe you could configure a separate SCOM subscription for Lync/Skype alerts. One that applies during the time when no maintenance tasks are running and one for the time when maintenance tasks are running. For the later one you could increase the value for the setting: Delay sending notifications if condition remains unchanged for longer then. However this only would work for alerts that automatically close.

I think the proper way is to place those servers into maintenance mode. you can easily schedule these maintenance mode using the SCOM Mantenance Mode Scheduler:


First, understand the difference between Alerts and notifications. Alerts are going to come and go, and aren’t necessarily what should drive you to action. You should, as previously stated in these responses, create a subscription to notify on alerts that you care about. You can build in “bake time” in the subscription that will delay emailing notification for a certain period, and then notify if the alert (monitor based alerts only) is still active after X minutes.

I agree with Tao, you should use Maintenance Mode, and the Maintenance Mode Scheduler is nice, but no longer free; it’s cheaper than Orchestrator! Also, consider only placing the objects that alert in maintenance mode, and not the entire computer object. This will ensure that you’re still monitoring the parts of the system that aren’t undergoing maintenance and be alerted/notified when those components have issues during the Lync/Skype maintenance.

On the horizon…and hopefully it stays, we should be able to schedule maintenance mode in the next version of OpsMgr 2016.

Tao has the best answer. I’d just add that you don’t have to put the whole server in maintenance mode if there is just one hosted instance that’s generating alerts. Just put that instance into MM. That way you don’t lose alerting on the operating system should it go down.

Also maintenance mode can be scheduled using the task scheduler and PowerShell. No additional software needed. I hear that OpsMgr 2016 will finally be able to schedule maintenance mode itself, hopefully with a GUI, but at least with PowerShell. MS may be all OMS all the time these days, but at least they are throwing us that bone.

Anyway, here’s a PowerShell (long) one-liner for setting a server into maintenance mode for one hour. Just change the class name, instance name, and end time to whatever works best in your situation.

Get-SCOMClass -DisplayName “Windows Computer” | Get-SCOMClassInstance | Where {$_.DisplayName -ieq “”} | Start-SCOMMaintenanceMode -EndTime (Get-Date).AddHours(1) -Reason PlannedApplicationMaintenance -Comment “Daily Scheduled Maintenance”

For the task scheduler, you can invoke this powershell code like this.

Program: C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe

Arguments: “& Get-SCOMClass -DisplayName ‘Windows Computer’ | Get-SCOMClassInstance | Where {$_.DisplayName -ieq ‘’} | Start-SCOMMaintenanceMode -EndTime (Get-Date).AddHours(1) -Reason PlannedApplicationMaintenance -Comment ‘Daily Scheduled Maintenance’”


What kind of alerts do you get? We have for example setup so that we set certain monitors in maintenance mode while windows update is running. Since the servers alert when communication is lost between members of a pool.

When it comes to tuning you have to make overrides yourself to make the monitors fit your needs. And when it comes to certain number of alerts you have to create your own monitor to handle that.