This is barely a PoC.

TLDR:
To simplify user experience, especially for the local dumper/ingestor, we should write a metadata.json file that contains the metadata of a specific dump.

Why:
Stuffing data in the filepath as some limitation:

• There's a limit of number of element in the data path (limited by the number of characters for each filename)
• Force the user to never rename the archive, which can be an issue when storing historical dumps (currently, the latest dump is computed by the file last modified time)
• Parsing the data can become more cumbersome than simply parsing a json file containing each value in the correct format.

How:
Add a metadata.json file that contains all the metadata around a specific dump.
It will then be reused by the ingestor to populate the RunID, the ClusterName and any future info we want.