If not go listen to this podcast – http://runasradio.com/default.aspx?showNum=81. If you think you know perfmon well, go listen to it anyway.
There’s also a video and an article linked in the show’s notes which are referred to several times in the podcast.
Top things I learned from that:
- There are 25 counters that all admins should be monitoring, regardless of what the server does
- Some counters are no longer useful with changes to the way disks are attached
- It is possible, and relatively easy, to have a blackbox perfmon trace permanently running and automatically starting on server start.
- It’s normal on Server 2008 to see some CPUs high and others not, because the cost of balancing the CPUs grows with the number of cores.