Adds Kafka based StateRepository implementation

[florian-stefan]

This PR adds a KafkaStateRepository. A detailed explanation of the functionality can be found in the Javadoc of the class KafkaStateRepository. It uses the standard Apache Kafka client library for connecting to Kafka as well as Gson serializing/deserializing feature states. Furthermore, the KafkaStateRepositoryTest uses the Kafka-JUnit4 test library from Salesforce for starting an in-memory Kafka cluster.

The file template.mf needs special attention because I was not really sure what to do here ... 😄.

Resolves: #304

[swen-fuhrmann]

I like the good documentation.
The logging is quite verbose on info level. I wonder if it make sense to log more on debug level.

[florian-stefan]

@smoczarski Thank you so much for your review! This is very helpful. I will add the suggested improvements as soon as possible!

[florian-stefan]

@smoczarski I agree that the logging was quite verbose: I decreased the log level of some logs to 'debug'.

[chkal]

Hi! Thanks a lot for creating this PR.

I just stepped down from my role as the Togglz maintainer and from being an active committer.

But I hope that some other @togglz/committer can help with the review.

[chkal]

I just mentioned your pull request in my email discussing the future of Togglz. I hope this will help to get more reviews. I'm really sorry for the delay, but you created this pull request in a very awkward period of the project. :-)

I don't know much about Kafka, so I'm not sure if my review is very helpful. However, there is one thing I would like to mention. Togglz is also creating a "distribution ZIP" file for people not using Maven. Could you perhaps include the new module in this distribution? Just have a look at this directory. It should be easy to understand if you look at the pom and the assembly descriptor.

Thanks a lot!