Added Kafka module to spark configuration to support Kafka integration.

[ilganeli]

This references issue:
https://github.com/NFLabs/zeppelin/issues/189

on the old Zeppelin GitHub.

This change lets us use Kafka with Spark streaming inside the shell.

[bzz]

Thank you for the contribution!

I have a question - is there a reason not to use %dep interpreter to get the jar only for the notebooks that require Kafka, at runtime, instead of packaging it with Zeppelin itself?

[jongyoul]

I've heard there isn't a graceful way to stop Spark streaming job on Zeppelin. Do you have any idea to do it?

[sejunra]

Any updates on this? Thanks.