We should have a better introspection for what is going on in pools: - how long does it take to start VMs - what is the startup success rate On top of this we can configure Nagios checks