We have an enterprise monitoring package. It has a module that plugs into VirtualCenter. It basically collects all the information avaiable to VirtualCenter. From there I can configure alerting / trending/ QoS data and such. I would like to configure the monitoring software so that I'm alerted to certain issues. I've got it setup for ESX server memory and cpu threshholds and virtual machine on/off states. That is fairly straight forward. It also collects all the disk information like Disk command issued, Disk write requests, Disk Bus resets, etc.. I don't really know how to determine what are values that I should be concerned about for these settings, specifically the Disk settings, like how many Disk command issued indicate poor performance and such. I'm sure some will depend on the equipment that I have but there should be some general guidelines for some of the settings and some rough estimates of what might indicate performance issues with disk.
Does anybody have any guidance or advice to offer.
You can use stress tools to measure your hardware power, and than you can stabilish your own guidelines for each server.
This is the most trustable way to acomplish this.
You can take a look on Microsoft web site for this tool, usually they are free.