clarify doc about default partitioner

[mdespriee]

Can I kindly suggest to add a very big warning in bold chars in the README and in --help about the partitioner used by default ?

I just spent entire DAYS trying to understand why my kafka-stream topology failed in some tests scenarios (test built around kafkacat), until I understand that what is a "default partitioner" is not a standard shared by all Kafka tools. Expecting data to be produced with java-like partitioner is, I think, not a crazy assumption. I don't think this kind of fine-print is enough.

Related to #131 obvisously