Determine whether to continue using Samza

[rhauch]

The prototype identified some shortcomings in Samza, and in particular the difficulty of deploying to OpenShift. While not a problem for the prototype, this would be a major requirement for any non-prototype project.

The prototype used on a few features of Samza, including the local database storage and simpler consumer API. But these were enough to make things quite a bit simpler for the prototype. So either these need to be reimplemented so Samza is not used, or it may be possible to add support for OpenShift to Samza. (It's not clear which is more effort.)