mon: support min_down_reporter conuted by subtree level

mon: support min_down_reporter conuted by subtree level

In many case OSDs in an isolated(public/cluster connection lost but osd<->mon is good) node
will report other OSD down to monitor,which usually wrongly mark someone down.

Nowaday the "osd_min_down_reporters", we would like to extend the semantic to
allow it counted by host or rack, thus user could require failure reports from at
least two nodes to mark an OSD down, which shoudl prevent an isoloated host make trouble
to the cluster.

Signed-off-by: Xiaoxi Chen <xiaoxi.chen@intel.com>