Description of problem:
Cluster has down OSD and is experiencing data rebalance
 - Shows negative value for disk IOPs and disk Throughput
 - Once down OSD returned to service (but still rebalancing), disk IOPs and Throughput show normal again

Version-Release number of selected component (if applicable):
ceph-ansible-3.0.31-1.el7cp.noarch                          Mon Apr 30 11:38:59 2018
ceph-common-12.2.4-6.el7cp.x86_64                           Mon Apr 30 11:38:47 2018
ceph-fuse-12.2.4-6.el7cp.x86_64                             Mon Apr 30 11:38:49 2018
ceph-installer-1.3.0-1.el7scon.noarch                       Tue Jul  4 09:31:42 2017
libcephfs2-12.2.4-6.el7cp.x86_64                            Mon Apr 30 11:37:29 2018
librados2-12.2.4-6.el7cp.x86_64                             Mon Apr 30 11:37:27 2018
libradosstriper1-12.2.4-6.el7cp.x86_64                      Mon Apr 30 11:37:30 2018
librbd1-12.2.4-6.el7cp.x86_64                               Mon Apr 30 11:37:27 2018
python-cephfs-12.2.4-6.el7cp.x86_64                         Mon Apr 30 11:37:29 2018
python-rados-12.2.4-6.el7cp.x86_64                          Mon Apr 30 11:37:27 2018
python-rbd-12.2.4-6.el7cp.x86_64                            Mon Apr 30 11:37:29 2018

cephmetrics-1.0-8.el7cp.x86_64                              Tue Mar 13 15:38:03 2018
cephmetrics-ansible-1.0-8.el7cp.x86_64                      Tue Mar 13 15:38:04 2018
cephmetrics-grafana-plugins-1.0-8.el7cp.x86_64              Tue Mar 13 15:37:59 2018


How reproducible:
First observation

Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info: