Add support for continuous archiving

[mbi]

A master-slave cluster configuration doesn't protect against potential data loss when e.g. the master (accidentally) deletes data and the loss is pushed to the slaves.

A solution such as wal-e continually pushes WAL archives to e.g. S3, so that one can always restore the cluster to any previous moment in time (as opposed to e.g. hourly backups)

Please consider adding support for continuous archiving via wal-e or an equivalent solution.