Well, I would like to clarify about first question:
does it mean that I cannot provide all VMs on one ESXi-host more than 4096 Virtual CPUs (config maximum)?
If so, I still can provide all VMs on that ESXi-host less than 4096 vCPUs even if my ESXi-host has only 24 logical CPUs?
About second question: according to configuration maximums document one ESXi-cluster can service 4000 VMs at maximum. But I don't understand WHY?
And I want do know it. What is a problem for cluster to service more than 4000 VMs?
Where is exist a bottleneck for maintaining? In cluster Database? And so on.