Ideally the metric should be in milliseconds instead of percentage (as shown in VMware
vCenter performance charts) to represent the actual time that the VM waits until the CPU
resources as allocated.
It would be great if it was added to Engine Reports as well.
Even a CLI script to convert percentage to ms would be sufficient in order to use the
metric in external tools (e.g. for alerting).
________________________________________
From: Michal Skrivanek <mskrivan(a)redhat.com>
Sent: Thursday, November 5, 2015 3:10 PM
To: RH - Itamar Heim
Cc: s k; users(a)ovirt.org; Doron Fediuck
Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning
On 05 Nov 2015, at 01:06, Itamar Heim <iheim(a)redhat.com>
wrote:
On 09/15/2014 02:54 PM, s k wrote:
> On 12/09/14 09:22, Itamar Heim wrote:
>> /With oVirt 3.5 nearing GA, time to ask for "what do you want to see in
/>/oVirt 3.6"? />//>/Thanks, />/Itamar
/>/_______________________________________________ />/Users mailing list />/Users
at
ovirt.org <
http://lists.ovirt.org/mailman/listinfo/users>
/>/http://lists.ovirt.org/mailman/listinfo/users />//>
>
>
> Performance metric similar to what VMware calls "CPU Ready" would be very
useful if it was available in the VM details in the admin portal.
>
>
> It would provide great visibility on VM's performance in an environment with CPU
overallocation.
>
I don't think we covered this in 3.6 (doron/michal)?
it should correspond to CPU steal time as reported by the host kernel
we do not report it up to the engine
It would be fairly trivial to add, however the interpretation of this value may be tricky