On Sun, Mar 8, 2020 at 5:13 PM M Ranga Swami Reddy <swamireddy(a)gmail.com> wrote:
Iam using the Luminous 12.2.11 version with prometheus.
On Sun, Mar 8, 2020 at 12:28 PM XuYun <yunxu(a)me.com> wrote:
> You can enable prometheus module of mgr if you are running Nautilus.
>
> > 2020年3月8日 上午2:15,M Ranga Swami Reddy <swamireddy(a)gmail.com> 写道:
> >
> > On Fri, Mar 6, 2020 at 1:06 AM M Ranga Swami Reddy <swamireddy(a)gmail.com
> >
> > wrote:
> >
> >> Hello,
> >> Can we get the IOPs of any rbd image/volume?
> >>
> >> For ex: I have created volumes via OpenStack Cinder. Want to know
> >> the IOPs of these volumes.
> >>
> >> In general - we can get pool stats, but not seen the per volumes stats.
> >>
> >> Any hint here? Appreciated.
Per image/volume stats are available since nautilus [1]. There is
no automated way to do it in luminous. Since you are using openstack,
you can probably set up some generic I/O monitoring at instance level
(i.e. external to ceph) and possibly apply disk I/O limits (again at
instance level, enforced in qemu).
[1]
https://ceph.io/rbd/new-in-nautilus-rbd-performance-monitoring/
Thanks,
Ilya