There are best practices out there from VMware and probably from storage vendors on how many VMs you should put on a particular LUN, this number is usually 10-20. We've gone ahead and defined standard sizes which is also useful for capacity manamgent so you know exactly how many shells you have at any given point. In terms of automatated tools, not really. What I've done is actually periodically poll the ESX Hosts from specific clusters and update a webpage on the disk capacity and number of VMs currently hosted on a LUN and based on our standard shells, the number of VMs that can possibly go on there based on a standard size. You can do something similiar and check the page to see which LUNs may be over prescribing. You definitely want to leave some room 20-30% for snapshots and growth. There is also a script out there that currently multipaths your storage routes so that you have an active/passive setup, but it'll peridocially change the path to even out the paths it uses to get to storage based on the number of blocks transfered or time duration.
I would also be interested to see if others have any other suggestions, but as far as I Know, is this going to be a manual process and will require some bookeeping if you want to stay within your threshold limits without overtaxing a LUN with contention.
