Currently getting number of VM via "fastline" is unlimited. It can lead to a situation were the fastline "steals" all VMs and normal integration do not proceed at all.

There should be a limit for how many VMs can be obtained via fastline.
The limit should be separate for MCU, packaging and RTA (other jenkins tasks?).
If the limit has been reached, the VM should be queued until there is capacity to allow it on the fastlane.
Limits should take into account different resources: Intel Macs, Arm Macs 11 and 12.

Open item: Should the limit be % of the full capacity or number of VMs?