We still use the old 'id' label for getting the number of OSDs in the
ceph cluster dashboard. However, the label was superseded by the
'ceph_daemon' label in one of the updates to the prometheus exporter in
ceph-mgr. This patch changes the query to what we use in other
dashboards -- i.e. 'ceph_daemon' instead of 'id'.
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1627725 Signed-off-by: Boris Ranto <branto@redhat.com>