I can't seem to find any documentation or best practice that would give any metrics to how many vm's on a single host can initiate a power on. I am planning to script this process so I want to make sure I don't cause problems by turning all vm's up to quickly. Any documentation or best practices around this?
I think a good best practice is to look at the number of slots that are available.
Alan did a great post on that, see HA Slot Size Information
Blog: lucd.info Twitter: @LucD22 Co-author PowerCLI Reference
Its maybe also worth to mind I/O Pressure for PowerOn Process.
if powering on with Bulk script - possible could Cause Disk Issues.
I think this is nothing globally to answer - as its very differently for each Infrastructure.
I Would just test it. Starting with 10 then increasing.
next to it analyzing cpu,mem and Disk.
Then you will get youre Value soon/later which fitting for youre Environment.
Ha Slots overally maybe not 100% reliable, as not covering each Layer
best regards
max