Throw parse exceptions on schema get errors for SchemaRegistryBasedAvroBytesDecoder

[jon-wei]

This PR adjusts SchemaRegistryBasedAvroBytesDecoder to throw ParseException instead of RE when failure to retrieve a schema occurs.

Throwing an RE can be problematic when this decoder is used in streaming ingestion: if a corrupt record enters the stream with an invalid schema ID, the user cannot easily recover from this as the streaming ingestion tasks will continue to fail and attempt to read the corrupt record.

This PR has:

• been self-reviewed.
• added documentation for new or modified features or behaviors.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[clintropolis]

i think this change makes sense 👍

my intentions were good when i made this change originally - a bad configuration non-intuitively seemed to run tasks that just couldn't process the data and threw everything out, so if the configuration would never work it made sense to fail fast.

However, I didn't consider the case of a bad record, where throwing a RE means there is no way to process such streams without skipping the offset past such a record, while a parse exception handles this case nicely.