This significantly simplifies cluster state handling by removing most of
the superfluous get_list() calls, replacing them with basic child reads
since most of them are just for a count anyways. The ones that require
states simplify this down to a child read plus direct reads for the
exact items required while leveraging the new read_many() function.