Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Whats the overhead on the client that runs and publishes data for the Systems Center Operations Manager? Most of the data that I want are the perfmon counters. In general, I don't worry to much about the performance overhead for the perfmon counters and will collect data on them about once a minute.

That said, I have no idea what the overhead on the management client is. Is it comparable to perfmon? Or should I be thinking about things completely differently?

share|improve this question
I don't have actual numbers on the overhead, I ran SCOM on 20 or so machines without issue. –  Joseph Kern Oct 22 '11 at 21:01
Does this answer your question? –  Bernie White Jun 2 '12 at 23:36

2 Answers 2

The SCOM sample frequency for performance counters is less then perfmon with most performance counter at a single sample every 5 - 15 minutes. High resolution counters will sample a few times a minute. That said SCOM would more likely be configured to collect a larger cross section of counters for baselining.

I would allow for an overhead of 5-10% CPU time. As there can be a lot things happening at one time. However the impact can be completely tuned to exactly what you want. Everything can be disabled or enabled as required.

Running a large environment of 500+ servers I would say that the impact is insignificant but it completely depends on the implementation.

SCOM server side it's a different question and there are calculator tools for capacity management.

Hope that helps.

share|improve this answer

The impact differs greatly based on a few factors:

  • Is your environment patched? Especially WMI an cscript versions can have an impact on resource utilization. Both windows 2003 and 2008 suffer a lot from this.
  • Is your datacenter well size? I have seen issues with clients overcommitting their datacenter greatly.
  • What management packs are running?

Generally speaking, yoru overhead should average around less than 5% CPU, with some spikes to 20% or even 50% when clearing the agent configuration. These spikes though are not happening a lot. Memory is usually a couple of 100 MB.

When activating SCOM APM, the impact can go as high as 20%, though again this depends on what application you monitor and how much data you want to collect.

Generally, the impact is neglegible, except of course if your datacenter is already undersized or you are running resource-sensitive management packs.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.