Usability/Design BUG: deployments/instances are never destroyed

[prudhomm]

This is a developer usability bug: nobody in practice wants to spend time undeploying and destroying apps. This is too cumbersome.
The consequence is that it is next to impossible to deploy a new blueprint, we need to undeploy/destroy instance from the cloudify interface.
I don't think we would like to continue that way and I believe it might be a design issue here.
I understand that the pipeline helps with re-running apps but we cannot reconfigure them anyway, so I am unsure about the usefulness of the current design.

[prudhomm]

@emepetres we discussed with @victorsndvg about some possible fixes for this. The click and run would definitely help

[emepetres]

#96 solved this issue partially: Deployments are destroyed, but their instances persist in order for the user to review "old" execution logs.

For deploying a new blueprint (update it), still some work must be done to allow updating it while not running instances exists in the system.