For emergencies, I think I would put more trust in a low-tech wood burner, moving extra wood in-house whenever there’s a serious storm forecast.

Still, in general it seems like the ultimate design constraint is always low probability events adding up - eg solar generation getting severely cut for a week, due to snow on the panels and cloudy days or hardware failure, coupled with many cold nights in a row. For which a fall back to a more traditional denser energy source (wood or gasoline) that you'd end up using every few years comes in much cheaper than trying to oversize the primary battery storage to handle every long tail event.