tbh thinking it's likely @digitalocean tho (hard crash), because have been using said @rethinkdb cluster with 0 (!) failures for ~1 year now