Collect only Performance data from SCOM Agent

Hi there,

has anyone here an idea what would be the best way to collect performance data from an agent without monitoring it? I need to collect Performance data from several systems but monitoring/alerting is not neccessary. My idea is to install the SCOM agent on these systems, put them together in a group and then disable all monitors for all objects in the group. Maybe someone came acroos this issue earlier…

That’s basically it - You need to disable all monitors for a group and drop objects in as needed.


It’ll be an interesting task moving forward, as any new management packs that get imported, will also need their monitors disabled.

Upgrade the agents to support Log Analytics. Add them to OMS/LA, add the perf counters you are looking for. Leave the health monitoring to SCOM, and then capture performance only data through that cloud thing Microsoft’s always talking about.

My thoughts:

  1. What performance data do you need? Check to see if there are data collection in rules in SCOM already or what you might need to custom author. Is it just OS metrics? Or SQL metrics?
  2. You can't just group servers \ computers together and override all the monitors for the group. This is because monitors won't just be targeted at servers \ computers but at other classes. And these classes won't necessarily roll up to a server or computer object. You'll need to create different different groups of objects and override each appropriately.
  3. There is an override creator tool that will help create the overrides - override-creator-for-scom-2012 -
I would ignore OMS on a variety of grounds:
  1. It is cloud based. Is Azure part of your organisations cloud strategy?
  2. Have your security team signed off on using cloud solutions? Are you allowed to have your servers directly connected to the Cloud? To avoid the latter you can use OMS gateways but this then requires deploying additional infrastructure.
  3. Do you really want to deploy and manage another toolset and pay for it when you already have something to hand that will work. Because deploying OMS will involve deploying new infrastructure (OMS gateways), firewall change requests, configuration.
  4. Who is paying for OMS? The free tier is 500 MB per day retained for 7 days. After that you start to get into allowances \ overage and the potential for costs to escalate really quickly.
Pragmatically (depending on number of servers \ number of alerts) you might find it easiest just to allow the alerting and configure overrides as you go along. It depends whether you end up with one false positive a week or one false positive every ten minutes.



This is not an option for me :slight_smile: