We have submitted a session on the VM and Baremetal platform for operators. This looks at another key set of users, those who use the APIs to write applications for the VM and BM platform.
We have historically found it hard to get feedback loops between the developers and API users of the Keystone, Nova, Neutron and Cinder APIs. The good news is the Heat, Trove and Magnum projects (to mention just a few) are members of our community great users of our APIs that are close to us. In a related way infra and the shade library work closely with our APIs.
Consider the VM and Baremetal platform giving controlled access to all your infrastructure (Storage, Networking, Compute (VM and baremetal)). Operators have invested in managing quotas, charge back, billing, etc. We should be able to make use of all that when running higher level services on that infrastructure such as: containers, databases, workflows etc. There is more discussion on this topic here: https://review.openstack.org/#/c/447031/
Lets get general feedback on the UX for API users and agree priorities around the most important things to get sorted:
* improved API docs, and common patterns
* improve API interoperability (find my external ip address, common patterns around bootstrapping a snapshot from a base image)
* improved interop testing to match the use cases these users need
* fine grained policy support
* per server keystone users (share barbican secrets with instances)
* services like trove using infrastructure that consumes from the users quota, but it is protected from accidental modification or deletion.
