Open Stack

On 8/15/2012 8:13 AM, Mark McLoughlin wrote:
> On Wed, 2012-08-15 at 12:52 +0100, Paul Bourke wrote:
>> Hi all,
>>
>> Swift contains some middleware called "healthcheck", with which you may or may
>> not be familiar.  It sits at the front of the pipeline and responds to GETs
>> with 200 under normal circumstances.
>>
>> I have submitted the following blueprint and patchset that ports this
>> functionality to openstack-common for inclusion across multiple Openstack
>> components:
>>
>> https://blueprints.launchpad.net/openstack-common/+spec/port-swift-healthcheck
>> https://review.openstack.org/#/c/11074/
>>
>> There has been some uncertainty as to whether this healthcheck is more valuable
>> than a simple GET/HEAD on the top level service endpoint.  While this could be
>> used, but it may be more consistent to have something which can be checked in
>> the same was across services.
> My own take on this is that consistently supporting HEAD on all our
> toplevel endpoints is much more obvious to API users than a /healthcheck
> resource.
>
> Cheers,
> Mark.
>
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Having each service provide a healthcheck was an idea discussed by 
operators at the last summit and would be a really good thing.  The 
intent was that monitoring tools would use it and that the healthcheck 
API should do whatever inspection the architects of the system believe 
is suitable to determine health. The architects know better than 
operators how to poke a system to determine if it is healthy.  Of course 
the posted code is not useful for this purpose and work would have to be 
done by each project to support such a common API.

  -David