Open Stack

>From: Jay Pipes [jaypipes at gmail.com] Wednesday, November 26, 2014 5:22 PM
>On 11/20/2014 08:12 AM, Sandy Walsh wrote:
>>
>> We're early in thinking about these problems, but would like to start the conversation now to get your opinions.
>>
>
>OK, so the goal of this effort should be to have a single OpenStack
>standard for what the payload and structure of notification messages
>will look like. That means, to me at least, that these schema files should:
>
>  a) Live in a single repo in the openstack/ code namespace
>(openstack/notification-schemas?)
>
>  b) Be published to an openstack.org subdomain, served by some static
>web server for all the world to read and/or mirror
>
>Let clients and servers that need to read and write these messages
>download the schemas as-needed.

Yep ... here's our current thinking. In order to keep the schemas close to the code that generates the notifications they would live in-tree with that service.

In the CI infrastructure, on a successful merge, a script would take the schema files and publish them to a well known location. This could be a tarball or a *.openstack.org webserver as you mention. 

Are you suggesting that all outside consumers would hit the openstack.org server to pull the schemas? That seems like a lot to impose on our infrastructure. We were thinking each service API would expose their schema via a new /schema resource (or something). Nova would expose its schema. Glance its own. etc. This would also work well for installations still using older deployments. 

That cool?

-S