feat(ingest): unbundle airflow plugin emitter dependencies

[cburroughs]

As described at length in
https://github.com/datahub-project/datahub/blob/v0.10.0/metadata-ingestion/setup.py#L62 the current state of platform compatibility for the confluent_kafka is complicated. It can be a source of installation incompatibility and general friction. To make the airflow plugin easier to include in resolvable dependencies sets across a variety of platforms, unbundle the datahub-kafka emitter into extras_require.

Checklist

• [X ] The PR conforms to DataHub's Contributing Guideline (particularly Commit Message Format)
• Links to related issues (if applicable)
• Tests for the changes have been added/updated (if applicable)
• [ X] Docs related to the changes have been added/updated (if applicable). If a new feature has been added a Usage Guide has been added for the same.
• [X ] For any breaking change/potential downtime/deprecation/big changes an entry has been made in Updating DataHub

[cburroughs]

(I think the slight asymmetry where the rest emitter is included by default instead of both being extras is the right tradeoff: For new people trying it out the rest emitter is almost certainly what they first want, and the use cases for excluding requests from an airflow install seem unlikely. I don't have a strong opinion however.)

[hsheth2]

LGTM

I pushed a small tweak to the docs

@cburroughs out of curiosity, what issues were you running into with confluent_kafka? Was it the same build issues on M1 macs or something else?

[cburroughs]

@cburroughs out of curiosity, what issues were you running into with confluent_kafka? Was it the same build issues on M1 macs or something else?

In short, yeah the lack of sufficient wheels for the versions within the datahub constraints.