Right now the out of the box thresholds in SRM are pretty much always wrong because they were set based on "typical" arrays from year ago. Several years ago VMware struggled with the same issue of not being able to tell what the capabilities of a connected array were so they worked with storage vendors to come up with their VASA protocol so the array can automatically tell VMware what it's capabilities are. So all the information we really need to set meaningful thresholds should already be available if we just requested it from VMware for our arrays. We could just set it to warn us at something like 80-90% of the VASA specs for an array at least for a more realistic starting point instead of the current manual scheme. Let's face it, nearly nobody sets those LUN and array performance thresholds accurately because it is entirely too tedious right now.