You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@stratos.apache.org by chris snow <ch...@gmail.com> on 2014/05/26 13:57:37 UTC

autoscaling - load average and memory consumption on docker/lxc instances

Hi Guys,

I noticed on my stratos+openstack+docker installation that stratos was
getting into a cycle trying to spawn new instances.

I think this is because load average (and probably memory consumption)
on lxc/docker instances is the same value as for the host [1].

Question: Am I correct in thinking that load average and memory
consumption autoscaling parameters aren't a good idea for docker
guests?

Many thanks,

Chris

---
[1] http://unix.stackexchange.com/questions/132021/is-a-lxc-docker-guest-load-average-the-same-as-the-hosts-load-average

Re: autoscaling - load average and memory consumption on docker/lxc instances

Posted by Shaheed Haque <sh...@cisco.com>.

On Saturday 31 May 2014 10:00:33 Lahiru Sandaruwan wrote:
> Hi Shaheed,
> 
> Could you explain the implementation a bit? Which technologies should be
> used in plug-able code, e.g. bash script etc.

Certainly. What we have in mind is a Java interface, maybe like this (ignore the exact 
names):

public class MonitorResult
{
    // range 0-100, notionally memory usage
    private double metric1;

    // range 0-100, notionally cpu usage
    private double metric2;

    public MonitorResult(double metric1, double metric2)
    {
...
    }
}

public interface IMonitorPlugin
{
    public boolean init();

    public void delete();

    public MonitorResult getMetrics();
        //throws Unobtainable;
}

The reason for Java and the init() and delete() hooks is to allow a .jar file to create - for 
example a TCP session or similar stateful object - and then use that to get the results 
each time.
     *  as hinted above, it is not clear to me what should happen on failure to get the 
values - suggestions welcome!
Bash could be supported by providing a Bash-calling .jar file. The selection of the plugin 
is via a System.property.

> Also can we extend this model to collect different stats user wants?

I considered that and concluded "yes and no":
     *  "Yes" in the sense that for pretty much any case I can think of, one value, 
normalised to the current 0-100% seems perfectly adequate.
         *   Given that we already have two, we may as well keep that (for the cases I cannot 
conceive of :-)).
             *  "No" in the sense that generalising support for more than two values seems 
redundant.
                 *  also support for N > 2 values seems like a LOT of work.

I hope that all sounds sane.

Thanks, Shaheed

> Thanks.
> 
> On Sat, May 31, 2014 at 1:51 AM, Shaheed Haque <sh...@cisco.com> wrote:
> > BTW, we are experimenting with a patch whose intent is to make the
> > memory/CPU
> > stats collection pluggable. We would definitely be interested in pushing
> > this
> > upstream - once we have something working.
> > 
> > On Wednesday 28 May 2014 22:37:32 Imesh Gunaratne wrote:
> > > Great! Thanks for the clarification Chris!
> > > 
> > > I just noted following from Docker's article on collecting metrics in
> > 
> > LXCs:
> > > "Docker will take care of this, or at least expose hooks to make it more
> > > straightforward. It is one of the reasons why we repeat over and over
> > > “Docker is not production ready yet”"
> > > 
> > > On Wed, May 28, 2014 at 9:01 PM, chris snow <ch...@gmail.com> wrote:
> > > > Thanks for the confirmation Nirmal.
> > > > 
> > > > On Wed, May 28, 2014 at 3:20 PM, Nirmal Fernando <
> > 
> > nirmal070125@gmail.com>
> > 
> > > > wrote:
> > > > > On Wed, May 28, 2014 at 4:49 PM, chris snow <ch...@gmail.com>
> > 
> > wrote:
> > > > >> Hi Imesh,
> > > > >> 
> > > > >> The autoscale policy relates to the cartridge, but on LXC you
> > > > >> aren't
> > > > >> monitoring the cartridge value, instead you are monitoring the
> > > > >> host.
> > > > >> If the load average hits a threshold because the host is presumably
> > > > >> under duress so the last thing you will want to do is spawn more
> > > > >> LXC
> > > > >> instances on the host?
> > > > > 
> > > > > Correct!
> > > > > 
> > > > >> Cheers,
> > > > >> 
> > > > >> Chris
> > > > >> 
> > > > >> 
> > > > >> 
> > > > >> On Wed, May 28, 2014 at 11:16 AM, Imesh Gunaratne <imesh@apache.org

Re: autoscaling - load average and memory consumption on docker/lxc instances

Posted by Imesh Gunaratne <im...@apache.org>.

Really nice to hear that Shaheed! Please create a Jira and attach the patch
once you are done.

Thanks


On Sat, May 31, 2014 at 1:51 AM, Shaheed Haque <sh...@cisco.com> wrote:

>
> BTW, we are experimenting with a patch whose intent is to make the
> memory/CPU
> stats collection pluggable. We would definitely be interested in pushing
> this
> upstream - once we have something working.
>
> On Wednesday 28 May 2014 22:37:32 Imesh Gunaratne wrote:
> > Great! Thanks for the clarification Chris!
> >
> > I just noted following from Docker's article on collecting metrics in
> LXCs:
> > "Docker will take care of this, or at least expose hooks to make it more
> > straightforward. It is one of the reasons why we repeat over and over
> > “Docker is not production ready yet”"
> >
> > On Wed, May 28, 2014 at 9:01 PM, chris snow <ch...@gmail.com> wrote:
> > > Thanks for the confirmation Nirmal.
> > >
> > > On Wed, May 28, 2014 at 3:20 PM, Nirmal Fernando <
> nirmal070125@gmail.com>
> > >
> > > wrote:
> > > > On Wed, May 28, 2014 at 4:49 PM, chris snow <ch...@gmail.com>
> wrote:
> > > >> Hi Imesh,
> > > >>
> > > >> The autoscale policy relates to the cartridge, but on LXC you aren't
> > > >> monitoring the cartridge value, instead you are monitoring the host.
> > > >> If the load average hits a threshold because the host is presumably
> > > >> under duress so the last thing you will want to do is spawn more LXC
> > > >> instances on the host?
> > > >
> > > > Correct!
> > > >
> > > >> Cheers,
> > > >>
> > > >> Chris
> > > >>
> > > >>
> > > >>
> > > >> On Wed, May 28, 2014 at 11:16 AM, Imesh Gunaratne <imesh@apache.org
> >
> > > >>
> > > >> wrote:
> > > >> > Hi Chris,
> > > >> >
> > > >> > Thanks for the quick response.
> > > >> > Right, I still don't get why we cannot adjust the threshold values
> > > >> > and
> > > >> > fix
> > > >> > it.
> > > >> >
> > > >> > Thanks
> > > >> >
> > > >> >
> > > >> > On Wed, May 28, 2014 at 11:37 AM, chris snow <chsnow123@gmail.com
> >
> > > >> >
> > > >> > wrote:
> > > >> >> Hi Imesh,
> > > >> >>
> > > >> >> I haven't looked into this in any more detail. All I have done is
> > > >> >> removed
> > > >> >> those parameters from my auto scaling policy.
> > > >> >>
> > > >> >> Many thanks,
> > > >> >> Chris
> > > >> >>
> > > >> >> On 28 May 2014 05:58, "Imesh Gunaratne" <im...@apache.org>
> wrote:
> > > >> >>> Hi Chris,
> > > >> >>>
> > > >> >>> Were you able to check the memory consumption and load average
> > >
> > > values
> > >
> > > >> >>> sent to Autoscaler by any chance?
> > > >> >>>
> > > >> >>> Thanks
> > > >> >>>
> > > >> >>>
> > > >> >>> On Mon, May 26, 2014 at 6:31 PM, chris snow <
> chsnow123@gmail.com>
> > > >> >>>
> > > >> >>> wrote:
> > > >> >>>> Hi Lahiru,
> > > >> >>>>
> > > >> >>>> It looks like there's some good info here [1].  However, as a
> > > >> >>>> first
> > > >> >>>> step, I was wondering whether we should add a note to the wiki
> to
> > > >> >>>> recommend not using the memory consumption or load average
> > >
> > > parameters
> > >
> > > >> >>>> for docker/lxc guests [2]?
> > > >> >>>>
> > > >> >>>> Cheers,
> > > >> >>>>
> > > >> >>>> Chris
> > > >> >>>>
> > > >> >>>> ---
> > > >> >>>> [1]
> > >
> > > http://blog.docker.io/2013/10/gathering-lxc-docker-containers-metrics/
> > >
> > > >> >>>> [2
> > >
> > >
> https://cwiki.apache.org/confluence/display/STRATOS/4.0.0+Auto-scaling+Gui
> > > de#id-4.0.0Auto-scalingGuide-memoryConsumptio>
> > > >> >>>> On Mon, May 26, 2014 at 1:24 PM, Lahiru Sandaruwan <
> > >
> > > lahirus@wso2.com>
> > >
> > > >> >>>> wrote:
> > > >> >>>> > Hi Chis.
> > > >> >>>> >
> > > >> >>>> >
> > > >> >>>> > On Mon, May 26, 2014 at 5:27 PM, chris snow <
> chsnow123@gmail.com
> > > >> >>>> >
> > > >> >>>> > wrote:
> > > >> >>>> >> Hi Guys,
> > > >> >>>> >>
> > > >> >>>> >> I noticed on my stratos+openstack+docker installation that
> > >
> > > stratos
> > >
> > > >> >>>> >> was
> > > >> >>>> >> getting into a cycle trying to spawn new instances.
> > > >> >>>> >>
> > > >> >>>> >> I think this is because load average (and probably memory
> > > >> >>>> >> consumption)
> > > >> >>>> >> on lxc/docker instances is the same value as for the host
> [1].
> > > >> >>>> >>
> > > >> >>>> >> Question: Am I correct in thinking that load average and
> memory
> > > >> >>>> >> consumption autoscaling parameters aren't a good idea for
> > > >> >>>> >> docker
> > > >> >>>> >> guests?
> > > >> >>>> >
> > > >> >>>> > Is there a way to find out the load average and memory
> > >
> > > consumption
> > >
> > > >> >>>> > of
> > > >> >>>> > the
> > > >> >>>> > guest?
> > > >> >>>> >
> > > >> >>>> > I think the proper solution is to customize/fix Cartridge
> agent
> > > >> >>>> > find
> > > >> >>>> > out
> > > >> >>>> > relevant values in these type of environments and publish
> stats
> > >
> > > to
> > >
> > > >> >>>> > CEP.
> > > >> >>>> >
> > > >> >>>> > Thanks.
> > > >> >>>> >
> > > >> >>>> >> Many thanks,
> > > >> >>>> >>
> > > >> >>>> >> Chris
> > > >> >>>> >>
> > > >> >>>> >> ---
> > > >> >>>> >> [1]
> > >
> > >
> http://unix.stackexchange.com/questions/132021/is-a-lxc-docker-guest-load->
> > average-the-same-as-the-hosts-load-average>
> > > >> >>>> > --
> > > >> >>>> > --
> > > >> >>>> > Lahiru Sandaruwan
> > > >> >>>> > Committer and PMC member, Apache Stratos,
> > > >> >>>> > Senior Software Engineer,
> > > >> >>>> > WSO2 Inc., http://wso2.com
> > > >> >>>> > lean.enterprise.middleware
> > > >> >>>> >
> > > >> >>>> > email: lahirus@wso2.com cell: (+94) 773 325 954
> > > >> >>>> > blog: http://lahiruwrites.blogspot.com/
> > > >> >>>> > twitter: http://twitter.com/lahirus
> > >
> > > >> >>>> > linked-in:
> > > http://lk.linkedin.com/pub/lahiru-sandaruwan/16/153/146
> > >
> > > >> >>>> --
> > > >> >>>> Check out my professional profile and connect with me on
> LinkedIn.
> > > >> >>>> http://lnkd.in/cw5k69
> > > >> >>>
> > > >> >>> --
> > > >> >>> Imesh Gunaratne
> > > >> >>>
> > > >> >>> Technical Lead, WSO2
> > > >> >>> Committer & PPMC Member, Apache Stratos
> > > >> >
> > > >> > --
> > > >> > Imesh Gunaratne
> > > >> >
> > > >> > Technical Lead, WSO2
> > > >> > Committer & PPMC Member, Apache Stratos
> > > >>
> > > >> --
> > > >> Check out my professional profile and connect with me on LinkedIn.
> > > >> http://lnkd.in/cw5k69
> > > >
> > > > --
> > > > Best Regards,
> > > > Nirmal
> > > >
> > > > Nirmal Fernando.
> > > > PPMC Member & Committer of Apache Stratos,
> > > >
> > > > Senior Software Engineer, WSO2 Inc.
> > > >
> > > > Blog: http://nirmalfdo.blogspot.com/
> > >
> > > --
> > > Check out my professional profile and connect with me on LinkedIn.
> > > http://lnkd.in/cw5k69
>
>


-- 
Imesh Gunaratne

Technical Lead, WSO2
Committer & PPMC Member, Apache Stratos

Re: autoscaling - load average and memory consumption on docker/lxc instances

Posted by Lahiru Sandaruwan <la...@wso2.com>.

Hi Shaheed,

Could you explain the implementation a bit? Which technologies should be
used in plug-able code, e.g. bash script etc.

Also can we extend this model to collect different stats user wants?

Thanks.


On Sat, May 31, 2014 at 1:51 AM, Shaheed Haque <sh...@cisco.com> wrote:

>
> BTW, we are experimenting with a patch whose intent is to make the
> memory/CPU
> stats collection pluggable. We would definitely be interested in pushing
> this
> upstream - once we have something working.
>
> On Wednesday 28 May 2014 22:37:32 Imesh Gunaratne wrote:
> > Great! Thanks for the clarification Chris!
> >
> > I just noted following from Docker's article on collecting metrics in
> LXCs:
> > "Docker will take care of this, or at least expose hooks to make it more
> > straightforward. It is one of the reasons why we repeat over and over
> > “Docker is not production ready yet”"
> >
> > On Wed, May 28, 2014 at 9:01 PM, chris snow <ch...@gmail.com> wrote:
> > > Thanks for the confirmation Nirmal.
> > >
> > > On Wed, May 28, 2014 at 3:20 PM, Nirmal Fernando <
> nirmal070125@gmail.com>
> > >
> > > wrote:
> > > > On Wed, May 28, 2014 at 4:49 PM, chris snow <ch...@gmail.com>
> wrote:
> > > >> Hi Imesh,
> > > >>
> > > >> The autoscale policy relates to the cartridge, but on LXC you aren't
> > > >> monitoring the cartridge value, instead you are monitoring the host.
> > > >> If the load average hits a threshold because the host is presumably
> > > >> under duress so the last thing you will want to do is spawn more LXC
> > > >> instances on the host?
> > > >
> > > > Correct!
> > > >
> > > >> Cheers,
> > > >>
> > > >> Chris
> > > >>
> > > >>
> > > >>
> > > >> On Wed, May 28, 2014 at 11:16 AM, Imesh Gunaratne <imesh@apache.org
> >
> > > >>
> > > >> wrote:
> > > >> > Hi Chris,
> > > >> >
> > > >> > Thanks for the quick response.
> > > >> > Right, I still don't get why we cannot adjust the threshold values
> > > >> > and
> > > >> > fix
> > > >> > it.
> > > >> >
> > > >> > Thanks
> > > >> >
> > > >> >
> > > >> > On Wed, May 28, 2014 at 11:37 AM, chris snow <chsnow123@gmail.com
> >
> > > >> >
> > > >> > wrote:
> > > >> >> Hi Imesh,
> > > >> >>
> > > >> >> I haven't looked into this in any more detail. All I have done is
> > > >> >> removed
> > > >> >> those parameters from my auto scaling policy.
> > > >> >>
> > > >> >> Many thanks,
> > > >> >> Chris
> > > >> >>
> > > >> >> On 28 May 2014 05:58, "Imesh Gunaratne" <im...@apache.org>
> wrote:
> > > >> >>> Hi Chris,
> > > >> >>>
> > > >> >>> Were you able to check the memory consumption and load average
> > >
> > > values
> > >
> > > >> >>> sent to Autoscaler by any chance?
> > > >> >>>
> > > >> >>> Thanks
> > > >> >>>
> > > >> >>>
> > > >> >>> On Mon, May 26, 2014 at 6:31 PM, chris snow <
> chsnow123@gmail.com>
> > > >> >>>
> > > >> >>> wrote:
> > > >> >>>> Hi Lahiru,
> > > >> >>>>
> > > >> >>>> It looks like there's some good info here [1].  However, as a
> > > >> >>>> first
> > > >> >>>> step, I was wondering whether we should add a note to the wiki
> to
> > > >> >>>> recommend not using the memory consumption or load average
> > >
> > > parameters
> > >
> > > >> >>>> for docker/lxc guests [2]?
> > > >> >>>>
> > > >> >>>> Cheers,
> > > >> >>>>
> > > >> >>>> Chris
> > > >> >>>>
> > > >> >>>> ---
> > > >> >>>> [1]
> > >
> > > http://blog.docker.io/2013/10/gathering-lxc-docker-containers-metrics/
> > >
> > > >> >>>> [2
> > >
> > >
> https://cwiki.apache.org/confluence/display/STRATOS/4.0.0+Auto-scaling+Gui
> > > de#id-4.0.0Auto-scalingGuide-memoryConsumptio>
> > > >> >>>> On Mon, May 26, 2014 at 1:24 PM, Lahiru Sandaruwan <
> > >
> > > lahirus@wso2.com>
> > >
> > > >> >>>> wrote:
> > > >> >>>> > Hi Chis.
> > > >> >>>> >
> > > >> >>>> >
> > > >> >>>> > On Mon, May 26, 2014 at 5:27 PM, chris snow <
> chsnow123@gmail.com
> > > >> >>>> >
> > > >> >>>> > wrote:
> > > >> >>>> >> Hi Guys,
> > > >> >>>> >>
> > > >> >>>> >> I noticed on my stratos+openstack+docker installation that
> > >
> > > stratos
> > >
> > > >> >>>> >> was
> > > >> >>>> >> getting into a cycle trying to spawn new instances.
> > > >> >>>> >>
> > > >> >>>> >> I think this is because load average (and probably memory
> > > >> >>>> >> consumption)
> > > >> >>>> >> on lxc/docker instances is the same value as for the host
> [1].
> > > >> >>>> >>
> > > >> >>>> >> Question: Am I correct in thinking that load average and
> memory
> > > >> >>>> >> consumption autoscaling parameters aren't a good idea for
> > > >> >>>> >> docker
> > > >> >>>> >> guests?
> > > >> >>>> >
> > > >> >>>> > Is there a way to find out the load average and memory
> > >
> > > consumption
> > >
> > > >> >>>> > of
> > > >> >>>> > the
> > > >> >>>> > guest?
> > > >> >>>> >
> > > >> >>>> > I think the proper solution is to customize/fix Cartridge
> agent
> > > >> >>>> > find
> > > >> >>>> > out
> > > >> >>>> > relevant values in these type of environments and publish
> stats
> > >
> > > to
> > >
> > > >> >>>> > CEP.
> > > >> >>>> >
> > > >> >>>> > Thanks.
> > > >> >>>> >
> > > >> >>>> >> Many thanks,
> > > >> >>>> >>
> > > >> >>>> >> Chris
> > > >> >>>> >>
> > > >> >>>> >> ---
> > > >> >>>> >> [1]
> > >
> > >
> http://unix.stackexchange.com/questions/132021/is-a-lxc-docker-guest-load->
> > average-the-same-as-the-hosts-load-average>
> > > >> >>>> > --
> > > >> >>>> > --
> > > >> >>>> > Lahiru Sandaruwan
> > > >> >>>> > Committer and PMC member, Apache Stratos,
> > > >> >>>> > Senior Software Engineer,
> > > >> >>>> > WSO2 Inc., http://wso2.com
> > > >> >>>> > lean.enterprise.middleware
> > > >> >>>> >
> > > >> >>>> > email: lahirus@wso2.com cell: (+94) 773 325 954
> > > >> >>>> > blog: http://lahiruwrites.blogspot.com/
> > > >> >>>> > twitter: http://twitter.com/lahirus
> > >
> > > >> >>>> > linked-in:
> > > http://lk.linkedin.com/pub/lahiru-sandaruwan/16/153/146
> > >
> > > >> >>>> --
> > > >> >>>> Check out my professional profile and connect with me on
> LinkedIn.
> > > >> >>>> http://lnkd.in/cw5k69
> > > >> >>>
> > > >> >>> --
> > > >> >>> Imesh Gunaratne
> > > >> >>>
> > > >> >>> Technical Lead, WSO2
> > > >> >>> Committer & PPMC Member, Apache Stratos
> > > >> >
> > > >> > --
> > > >> > Imesh Gunaratne
> > > >> >
> > > >> > Technical Lead, WSO2
> > > >> > Committer & PPMC Member, Apache Stratos
> > > >>
> > > >> --
> > > >> Check out my professional profile and connect with me on LinkedIn.
> > > >> http://lnkd.in/cw5k69
> > > >
> > > > --
> > > > Best Regards,
> > > > Nirmal
> > > >
> > > > Nirmal Fernando.
> > > > PPMC Member & Committer of Apache Stratos,
> > > >
> > > > Senior Software Engineer, WSO2 Inc.
> > > >
> > > > Blog: http://nirmalfdo.blogspot.com/
> > >
> > > --
> > > Check out my professional profile and connect with me on LinkedIn.
> > > http://lnkd.in/cw5k69
>
>


-- 
--
Lahiru Sandaruwan
Committer and PMC member, Apache Stratos,
Senior Software Engineer,
WSO2 Inc., http://wso2.com
lean.enterprise.middleware

email: lahirus@wso2.com cell: (+94) 773 325 954
blog: http://lahiruwrites.blogspot.com/
twitter: http://twitter.com/lahirus
linked-in: http://lk.linkedin.com/pub/lahiru-sandaruwan/16/153/146

Re: autoscaling - load average and memory consumption on docker/lxc instances

Posted by Shaheed Haque <sh...@cisco.com>.

BTW, we are experimenting with a patch whose intent is to make the memory/CPU 
stats collection pluggable. We would definitely be interested in pushing this 
upstream - once we have something working.

On Wednesday 28 May 2014 22:37:32 Imesh Gunaratne wrote:
> Great! Thanks for the clarification Chris!
> 
> I just noted following from Docker's article on collecting metrics in LXCs:
> "Docker will take care of this, or at least expose hooks to make it more
> straightforward. It is one of the reasons why we repeat over and over
> “Docker is not production ready yet”"
> 
> On Wed, May 28, 2014 at 9:01 PM, chris snow <ch...@gmail.com> wrote:
> > Thanks for the confirmation Nirmal.
> > 
> > On Wed, May 28, 2014 at 3:20 PM, Nirmal Fernando <ni...@gmail.com>
> > 
> > wrote:
> > > On Wed, May 28, 2014 at 4:49 PM, chris snow <ch...@gmail.com> wrote:
> > >> Hi Imesh,
> > >> 
> > >> The autoscale policy relates to the cartridge, but on LXC you aren't
> > >> monitoring the cartridge value, instead you are monitoring the host.
> > >> If the load average hits a threshold because the host is presumably
> > >> under duress so the last thing you will want to do is spawn more LXC
> > >> instances on the host?
> > > 
> > > Correct!
> > > 
> > >> Cheers,
> > >> 
> > >> Chris
> > >> 
> > >> 
> > >> 
> > >> On Wed, May 28, 2014 at 11:16 AM, Imesh Gunaratne <im...@apache.org>
> > >> 
> > >> wrote:
> > >> > Hi Chris,
> > >> > 
> > >> > Thanks for the quick response.
> > >> > Right, I still don't get why we cannot adjust the threshold values
> > >> > and
> > >> > fix
> > >> > it.
> > >> > 
> > >> > Thanks
> > >> > 
> > >> > 
> > >> > On Wed, May 28, 2014 at 11:37 AM, chris snow <ch...@gmail.com>
> > >> > 
> > >> > wrote:
> > >> >> Hi Imesh,
> > >> >> 
> > >> >> I haven't looked into this in any more detail. All I have done is
> > >> >> removed
> > >> >> those parameters from my auto scaling policy.
> > >> >> 
> > >> >> Many thanks,
> > >> >> Chris
> > >> >> 
> > >> >> On 28 May 2014 05:58, "Imesh Gunaratne" <im...@apache.org> wrote:
> > >> >>> Hi Chris,
> > >> >>> 
> > >> >>> Were you able to check the memory consumption and load average
> > 
> > values
> > 
> > >> >>> sent to Autoscaler by any chance?
> > >> >>> 
> > >> >>> Thanks
> > >> >>> 
> > >> >>> 
> > >> >>> On Mon, May 26, 2014 at 6:31 PM, chris snow <ch...@gmail.com>
> > >> >>> 
> > >> >>> wrote:
> > >> >>>> Hi Lahiru,
> > >> >>>> 
> > >> >>>> It looks like there's some good info here [1].  However, as a
> > >> >>>> first
> > >> >>>> step, I was wondering whether we should add a note to the wiki to
> > >> >>>> recommend not using the memory consumption or load average
> > 
> > parameters
> > 
> > >> >>>> for docker/lxc guests [2]?
> > >> >>>> 
> > >> >>>> Cheers,
> > >> >>>> 
> > >> >>>> Chris
> > >> >>>> 
> > >> >>>> ---
> > >> >>>> [1]
> > 
> > http://blog.docker.io/2013/10/gathering-lxc-docker-containers-metrics/
> > 
> > >> >>>> [2
> > 
> > https://cwiki.apache.org/confluence/display/STRATOS/4.0.0+Auto-scaling+Gui
> > de#id-4.0.0Auto-scalingGuide-memoryConsumptio> 
> > >> >>>> On Mon, May 26, 2014 at 1:24 PM, Lahiru Sandaruwan <
> > 
> > lahirus@wso2.com>
> > 
> > >> >>>> wrote:
> > >> >>>> > Hi Chis.
> > >> >>>> > 
> > >> >>>> > 
> > >> >>>> > On Mon, May 26, 2014 at 5:27 PM, chris snow <chsnow123@gmail.com
> > >> >>>> > 
> > >> >>>> > wrote:
> > >> >>>> >> Hi Guys,
> > >> >>>> >> 
> > >> >>>> >> I noticed on my stratos+openstack+docker installation that
> > 
> > stratos
> > 
> > >> >>>> >> was
> > >> >>>> >> getting into a cycle trying to spawn new instances.
> > >> >>>> >> 
> > >> >>>> >> I think this is because load average (and probably memory
> > >> >>>> >> consumption)
> > >> >>>> >> on lxc/docker instances is the same value as for the host [1].
> > >> >>>> >> 
> > >> >>>> >> Question: Am I correct in thinking that load average and memory
> > >> >>>> >> consumption autoscaling parameters aren't a good idea for
> > >> >>>> >> docker
> > >> >>>> >> guests?
> > >> >>>> > 
> > >> >>>> > Is there a way to find out the load average and memory
> > 
> > consumption
> > 
> > >> >>>> > of
> > >> >>>> > the
> > >> >>>> > guest?
> > >> >>>> > 
> > >> >>>> > I think the proper solution is to customize/fix Cartridge agent
> > >> >>>> > find
> > >> >>>> > out
> > >> >>>> > relevant values in these type of environments and publish stats
> > 
> > to
> > 
> > >> >>>> > CEP.
> > >> >>>> > 
> > >> >>>> > Thanks.
> > >> >>>> > 
> > >> >>>> >> Many thanks,
> > >> >>>> >> 
> > >> >>>> >> Chris
> > >> >>>> >> 
> > >> >>>> >> ---
> > >> >>>> >> [1]
> > 
> > http://unix.stackexchange.com/questions/132021/is-a-lxc-docker-guest-load-> > average-the-same-as-the-hosts-load-average> 
> > >> >>>> > --
> > >> >>>> > --
> > >> >>>> > Lahiru Sandaruwan
> > >> >>>> > Committer and PMC member, Apache Stratos,
> > >> >>>> > Senior Software Engineer,
> > >> >>>> > WSO2 Inc., http://wso2.com
> > >> >>>> > lean.enterprise.middleware
> > >> >>>> > 
> > >> >>>> > email: lahirus@wso2.com cell: (+94) 773 325 954
> > >> >>>> > blog: http://lahiruwrites.blogspot.com/
> > >> >>>> > twitter: http://twitter.com/lahirus
> > 
> > >> >>>> > linked-in:
> > http://lk.linkedin.com/pub/lahiru-sandaruwan/16/153/146
> > 
> > >> >>>> --
> > >> >>>> Check out my professional profile and connect with me on LinkedIn.
> > >> >>>> http://lnkd.in/cw5k69
> > >> >>> 
> > >> >>> --
> > >> >>> Imesh Gunaratne
> > >> >>> 
> > >> >>> Technical Lead, WSO2
> > >> >>> Committer & PPMC Member, Apache Stratos
> > >> > 
> > >> > --
> > >> > Imesh Gunaratne
> > >> > 
> > >> > Technical Lead, WSO2
> > >> > Committer & PPMC Member, Apache Stratos
> > >> 
> > >> --
> > >> Check out my professional profile and connect with me on LinkedIn.
> > >> http://lnkd.in/cw5k69
> > > 
> > > --
> > > Best Regards,
> > > Nirmal
> > > 
> > > Nirmal Fernando.
> > > PPMC Member & Committer of Apache Stratos,
> > > 
> > > Senior Software Engineer, WSO2 Inc.
> > > 
> > > Blog: http://nirmalfdo.blogspot.com/
> > 
> > --
> > Check out my professional profile and connect with me on LinkedIn.
> > http://lnkd.in/cw5k69