Open Stack

On 07/26/2017 03:28 PM, David Moreau Simard wrote:
> Hi,
> 
> A bit out of my comfort zone for the content of your giant pile of
> thoughts, but I'll jump on this bit:
> 
>> We have a desire to add something that knows how to
>> render the job-output.json in some nice ways.
> 
> The next things on my plate for ARA are:
> - Introduce the concept of API to abstract model access
> - Input drivers based on the API (callback, zuul json, mqtt, etc.)
> - Output drivers based on the API (html, junit, subunit, influx, graphite, etc.)
> 
> This is no small task but I'm hoping to have made significant progress
> by the time the PTG rolls around.
> 
> Paul showed interest in having those job-output.json files imported into ARA.
> I could see this taking a similar approach than how we are currently
> importing subunit files into openstack-health.

I actually may want to discuss doing this in the other direction - but I 
think that's a post-PTG discussion. Basically - the way we generate that 
json today is pretty inefficient (we have to load and rewrite it 
multiple times) The way ara collects data into a database then exports 
it would actually be MUCH nicer and involve less data in RAM at any one 
given point in time.

The downside is that there are a few 'non-standard' pieces of important 
metadata that we'd need to work out how to get ara to deal with without 
ara growing zuul-specific knowledge.

In any case - I think integration level is a great conversation topic 
for post-PTG. For now I think the level of ara integration we'll have 
just with the simple callback+export will be great and will set us up 
well to iterate and improve on it.

> I'll let you know if I have anything to show !
> 
> David Moreau Simard
> Senior Software Engineer | Openstack RDO
> 
> dmsimard = [irc, github, twitter]
> 
> _______________________________________________
> OpenStack-Infra mailing list
> OpenStack-Infra at lists.openstack.org
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
>