How to dynamically enable/disable virtual machines in the cloud depending on the load?

Question

Virtualization

Initial conditions:
there is a cloud service that provides services on the IaaS model.
this cloud has an api and a client for this api to manage virtual machines.
there is a terminal farm on Windows Server.
there is a monitoring system in the form of Zabbix
The essence of the problem:
I would like to turn on / off individual hosts, depending on the total load of the farm. Let's say there are 4 hosts for senas, first we turn on only one. Then, when resources almost run out on it, turn on the next one, etc. and, accordingly, vice versa, turn off the least loaded hosts.

We thought about taking metrics from zabbix (RAM consumption, but maybe someone else will tell you) with self-written software. Watch the average value for a given period and from this make a decision to turn on / off the host.

I tried to google, but somehow I didn't find anything particularly suitable. Can someone tell me there is a ready-made tool and we are trying to reinvent the wheel? ))))

PS
There was an option to turn off all hosts, except for one at the end of the working day, but the production cycle is round-the-clock, so I would like to automate more deeply =)

Answer 1

Yes, there is a finished product - Citrix VDI.
There is also Windows Virtual Desktop. This is only in the Azure cloud.