Fix: Retention for Distributed/Standalone

[nikhilsinhaparseable]

For Standalone -
Fix retention for standalone deployments
not working because of .stream folder created to store stream.json file

For Distributed -
Querier -
deletes parquet files from S3
sends retention info to ingesters' API (new)

Ingester -
New API to be created - PUT /logstream/{logstream}/UpdateOnRetention
On API call from querier - update snapshot, first_event_at in its own stream.json