Technically speaking fully-automatic live-backup-replication is not a backup at all (same reason why RAID is not a backup), so in that sense it makes sense to store the data in parallel, rather than replicating it forwards like a chain. It is actually on my long-term roadmap to support multiple simultaneous database/storage types, so I may as well adjust that a bit and support ambiguous database/storage types (ie. multiple InfluxDB) to support this use-case as well 
I’ll bump this feature forwards on my todo-list a little bit, but I don’t want to give any promises yet on when this will be done