Console needs to decouple from a specific schema

[nitisht]

Console needs to ensure any schema that user has sent to a stream is properly visible and understandable in the dashboard. Those changes are going on in the console repo. But there two specific things backend needs to ensure:

All the log data that is sent as body in POST request gets stored as is in the parquet file. But server should have two special columns in each stream (also stored in the parquet file).

1. p_metadata
2. p_tags

Both these fields are sent as HTTP headers during the post request. So server would parse any field starting with x-p-meta:m1=v1,m2=v2 to p_metadata and x-p-tags:t1=v1,t2=v2 to p_tags.

Also both these special fields should be returned in the get schema api call for a stream.