Add support for deploying stand alone Spark clusters

[razvan]

Description

Jupyter notebooks can connect to stand alone clusters. This allows for much better scalability when analyzing big data with Pyspark notebooks.

Acceptance

• A new custom resource describes stand alone clusters
• Stand alone clusters can be are provisioned with external dependencies
• Stand alone clusters can be connected with history servers
• ...

[sbernauer]

I personally like Notebooks running in client mode and requesting executors and would prefer that solution over a Spark standalone cluster. We used that extensively at my previous company