[Gluster-Maintainers] Community Transparency [Was: Responsibilities and expectations of our maintainers]

[Niels de Vos]

On Tue, Jun 23, 2015 at 11:13:06AM -0400, Vijay Bellur wrote:
> On Tuesday 23 June 2015 10:35 AM, Dan Lambright wrote:
> >
> >
> >----- Original Message -----
> >>From: "Vijay Bellur" <vbellur at redhat.com>
> >>To: maintainers at gluster.org
> >>Sent: Tuesday, June 23, 2015 9:43:11 AM
> >>Subject: Re: [Gluster-Maintainers] Community Transparency [Was: Responsibilities and expectations of our maintainers]
> >>
> >>Hey All,
> >>
> >>I know that all of us are quite occupied at this point in time. I would
> >>love to have some responses on this one. Will appreciate greatly if you
> >>can spare some time off your schedule and chime in with your inputs.
> >>
> >>Thanks,
> >>Vijay
> >>
> >>On Friday 19 June 2015 06:30 AM, Vijay Bellur wrote:
> >>>Re-kindling the discussion again on this list.
> >>>
> >>>I agree with everything that Niels mentions here as expectations. In
> >>>addition I would like us to start thinking about building more
> >>>transparency in the community on how we operate as maintainers and also
> >>>lead by example so that it is very evident to the broader community
> >>>about we being serious about walking the talk.
> >>>
> >>>I would like us to think about a small list so that it can be
> >>>accomplished sooner than later. My top 3 for this list would be:
> >>>
> >>>1. Some of the goals for maintainers are measurable: reviews completed,
> >>>reviews outstanding, attendance in IRC meetings, bugzilla backlog,
> >>>patches sent etc. We can send a report at some frequency
> >>>(bi-weekly/monthly) about these aspects so that the community observes
> >>>the quantitative aspects of our involvement.
> >
> >Can some of these items be automated?
> >
> 
> Yes, most of these can be automated. I already have scripts that pull data
> off gerrit. bugzilla can be easily automated too, for IRC meetings we have
> logs and hence pulling out statistics should not be difficult. If we agree
> on the set of quantifiable metrics that we'd like to see, I can work on
> making that happen.

Last week while doing booth duty at the DevOps Days, I've started to
look into using Elasticsearch/Logstash/Kibana for getting some
statistics on regression tests. Now I've setup an ELK stack for this, I
fed it some data from Gerrit too.

Kibana is the webui for accessing the data, searches, graphs and
dashboards. There is no read-only mode that I have found, so currently
anyone with access has full control. If you would like to have a look,
and maybe build some graphs yourself, you can pass me a username and a
password-hash generated with 'openssl passwd -crypt'.

Two screenshots with some ideas on what can get gathered:
  - https://devos.fedorapeople.org/tmp/gerrit-on-elk.png
  - https://devos.fedorapeople.org/tmp/patchcounts-per-developer.png

Cheers,
Niels