In Orchestration::Compute https://github.com/theforeman/foreman/blob/develop/app/models/concerns/orchestration/compute.rb#L152 we are waiting for default amount of time for machine to become ready. But for compute resources like Openstack we can do much better.

See https://docs.openstack.org/developer/nova/vmstates.html
Machine can easilly end in ERROR state (For example bad av zone or flavor av zone combination,it can take seconds to be in this state)
We are than waiting i think 300s (dafault wait_for timeout) to find out.
We should be able to speedup the feedback loop