Investigate Performance Optimization For Export Fetching

For this card, we want to investigate how much impact on memory will be use if we send export JSON as a stream to export service for upload.

Goal:
We want memory usage for export usage to be the same for both scenarios we want constant memory :
1. Single subscription 
2. Pick a large number 10,000 + subscriptions 

Requirements:

Look at the generated class to add the `existingJavaType` yaml field for the export model class :

Example: java.util.stream.Stream<org.candlepin.subscriptions.json.Subscription>
Keep in mind that we may have to refactor yaml to properly generate streams class for Json body 

• Leverage profiler tools to track how java Stream API would effect memory usage and ways to optimize implementation.
• Create a unit test to track the memory performance for streams of POJO's  with jsonschema2pojo leverage existing Types
• We want to use jackson to write to a tmp file and than upload file to export API client. May need to tweak here and to watch out for client causing memory to be consumed
• Use File as the body (and it does upload properly), seems to have a 50MB cap, we can probably live with that for the short term at least.

Testing:

Verify Memory usage performance using profiler for memory benchmark.
Do we want to bring perf (optional)? - Work with performance team to set up Continuous Performance Test (CPT) in this area. Yes: SWATCH-2123