How to configure retention of `*.csv.gz`, `*.parquet` and `*.chunk.parquet` files in GCP cloud storage?

**Problem**

How to cap the number of `*.csv.gz` files in cloud storage?

**Details**

We use the export bucket functionality with GCP cloud storage together with BigQuery. Cube instructs BigQuery to export buckets in to `*.csv.gz` whenever the refresh-scheduler recognizes a change. My question is how to handle retention of the `*.csv.gz` files?

- Are they needed long-term to restore caches of cubestore-workers, or are the `*.parquet` and `*.chunk.parquet` files used for this?
- If we can set a retention config, which one is recommended?
- Does cube manage retention?

Thank you!