EthRPC sqlite operational requirements

[GlacierWalrus]

I have a few queries about the operational requirements for the EthRPC sqlite databases.

1. Are the database growths bounded somehow? e.g. Will lotus perform housekeeping/garbage collection on the data to limit the growth? I see there's configuration for EthTxHashMappingLifetimeDays, but I don't see a similar config for the FEVM Events store
2. Do we have any projections of DB storage growth relative to EthRPC traffic?
3. In the event of a gateway failure, are API service providers expected to restore the sqlite database to a replacement instance?
4. If the answer to (3) is yes, do we need to maintain backups of the database to account for hardware failure?
5. Does lotus expose any metrics to monitor the state of the database, or do we need to rely solely on disk usage monitoring?