How many virtual resources are wasted in Test/Dev environments?

Today internal Test/Dev teams can easily have hundreds or thousands of engineers, using Cloud (internal or external) to write, build and test applications. Self-service is an important feature of a Cloud, but it also means that every developer can create VMs for their use.   

Such Test/Dev environments can easily get out of control and managing cost of a Test/Dev environment becomes an issue

      -  Can you tell how many VMs in the environment are doing useful job and which should be deleted?    
      -  Once you reach the limit of your cloud, is it time to buy more hardware, or should you perform clean up? How to identify VMs for cleanup?    
      -  Can your internal cloud handle a spike in a number of simultaneous VMs – for instance caused by a test cycle?

If your customer(s) have these or similar issues, we’d love to get your feedback, as we’re working on internal tool to address such issues for internal EMC environments and would like to explore its potential for wider applicability.