[Openstack-operators] Atlanta Summit - More Ops? ;)

Narayan Desai narayan.desai at gmail.com
Mon Mar 31 19:01:04 UTC 2014


On Mon, Mar 31, 2014 at 12:21 PM, matt <matt at nycresistor.com> wrote:

> Incentives is a very interesting question.
> What sort of incentives exist for devs?
> Acknowledged as Contributor ( ATC ) to OpenStack -- ( resume fodder, free
> ticket to summit )
> Voting in PTL, TC
> I guess those are the official benefits.  There are the unofficial ones
> such as being saught after in the job market.

I think that beyond those top level incentives, there are also more subtle
recognitions as well. Loads of metrics are derived from code contributions,
companies compete over it, etc. Openstack is primarily viewed by most
people as a coding project. (I think this is the wrong strategy, but I
think that the crux of the issue is this view today)

> What sort of incentives do you envision for operators?

I think that we should start by considering the ways that we want operators
to contribute:
 - design review
 - detailed bug reports
 - requirements* (this one i think is the best handled at the moment, but
could bear improvement)

Basically, you have people who understand how large systems need to work in
order to function well; you want to capture their expertise and use it to
improve the openstack design/implementation. Telling these people to commit
code is pretty much ass backward, both because they have a different kind
of role, and a different set of expertise, though it is what we tell them
to do today. (they can write doc, which helps others, but not the
contributor, at least not directly).

IMO, the big problem is that there isn't a receptor site for these
contributions. If you file tickets, no one pays attention. There is an
argument that this is the role of commercial outfits, but that doesn't
result in a particularly compete open effort. As far as I can tell, this is
because we don't give devs any incentive to pay attention to real
deployments, aside from any local relationships they might have.

(as an aside, the project has developed a great immune system to cope with
the extreme growth it has faced, but this results in an large barrier to
entry when ops folks try to casually contribute something)

I mean with ATCs it's easy to set a barrier... did you contribute code.
> How do you quantify an operators contribution to OpenStack?  When an
> operator is distinguished as an active contributor do we provide them with
> the same abilities as an ATC in dev?  There is no PTL for operators... no
> need for voting for TC... though maybe they should?

I think this activity should be focused around detailed critiques and bug
reports. That is *the* way that you get ops people involved. You could run
metrics in a similar fashion to the dev side.

I suspect that the early addition of this kind of feedback would result in
a usable ceilometer, just to pick an example out of the air.

> It raises interesting questions.  And I think it starts with, do we want
> to create a role for an ATC outside of development ( I think we already do
> for documentation contributors ).  If so, are they an ATC or an ATC*.  Then
> comes the question of quantifying contribution.
> What are your thoughts?

I'm not so concerned about the metrics/gamification of it. I'm just tired
of not having anyone to productively work with.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140331/102b742a/attachment-0001.html>

More information about the OpenStack-operators mailing list