[Starlingx-discuss] [Testing][Performance metrics] Feedback required

Perez, Ricardo O ricardo.o.perez at intel.com
Fri Aug 17 13:14:43 UTC 2018


> -----Original Message-----
> From: Chris Friesen [mailto:chris.friesen at windriver.com]
> Sent: Wednesday, August 15, 2018 6:28 PM
> To: starlingx-discuss at lists.starlingx.io
> Subject: Re: [Starlingx-discuss] [Testing][Performance metrics] Feedback
> required
> 
> On 08/15/2018 03:37 PM, Perez, Ricardo O wrote:
> > Hello StarlingXers,
> >
> > We have been thinking in how to measure StarlingX performance, in
> > order to do so, we would like to present you an initial proposal in
> > order to get feedback and ideas from you guys.
> >
> > The proposed metrics for StarlingX performance are:
> >
> >   * Detection of failed VM – tracked on milliseconds
> 
> Are we talking the qemu process or intrusive guest monitoring? (Or both?)
[Perez, Ricardo O] By now, we are talking about QEMU, but when bare metal became fully functional we can go through guest monitoring also.> 

> >   * Detection of failed compute node – tracked in milliseconds
> 
> Different failure modes (power outage, mgmt link failure, critical process
> failure) could result in different detection times.
[Perez, Ricardo O] Yes, I agree on that, and I would like to ask if you believe we should include all of them, or just some ? which ones could be more useful ? 
> 
> >   * Auto controller node failure recovery - No impact on StarlingX
> >   * Network link failure detection – tracked in milliseconds with no major
> >     impact on StarlingX
> 
> What about network failure beyond the first hop (so we don't lose carrier)?
[Perez, Ricardo O] That sounds good, in fact, the idea is to test the failure between source / destination, mostly between VM to VM, Compute to Controller. The first hop you mean between Carrier (Internet) and Controller ? or something else ? 
 
> 
> How about time from dead-office-recovery to instance network connectivity?
[Perez, Ricardo O] You mean from a complete power off until we become available to the network ? and it would be for the controller ?, VM's ? , other scenario ? 
> 
> Do we want to consider OpenStack performance?
[Perez, Ricardo O] Maybe, but as we are taking the Open Stack and tunning some areas, I think we should focus only in the areas that we are touching, instead of consider "all" OpenStack performance, what do you think ? 
> 
> 
> Chris
[Perez, Ricardo O] Thanks for all your comments Chris :)
> 
> _______________________________________________
> Starlingx-discuss mailing list
> Starlingx-discuss at lists.starlingx.io
> http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss


More information about the Starlingx-discuss mailing list