I have several ESXi instances in my environment with VMs on it which I need only for testing. I have also several jobs which triggers the tests. because I have very different technologies and because I do not want to maintain all VMs with all technologies I have labeled the VMs and also the jobs. So some jobs can be executed everywhere, some others only on specific VMs.

The problem is that a waiting job in the queue triggers the deployment of all VMs currently not in use. If the job cannot be executed on these VMs a loop of deploying and undeploying VMs will be triggered until the job will be handled by an executor.

Suggestion for a solution:
Before a VM gets deployed it should be checked if the labels will allow an execution on this executor.