We should implement this write up for a small cut out of encrytped storage for each peer: lightning/bolts#1110. This helps for the case where maybe you lost everything but your seed, and you also had mainly/all public channels. Once you boot your node back up, you can download gossip, then go seek out those peers to give you your encrypted backup.

We should be able to use the SCB scheme as is, since it's already encrypted with an AEAD scheme. On the implementation side, we should take care to ensure that each peer only has as a fixed amount of storage allocated, that these are cleaned up on a delay after all channels are closed, and it might be useful to have some RPCs to fetch/store this data on demand.

The SCB flow should also be updated to attempt to fall back to this method, if we start p in recovery mode with a partial SCB (would need a new explicit path).