Hi Tomek Thanks for the feedback. I've included some comments inline.
On 2021/05/12 08:37:56, Tomasz Urbaszek <turbas...@apache.org> wrote: > > 1) How far do we support Kibble-1 while the new Kibble is being developed? > > To be honest I think we as a project don't have enough capacity to > both maintain old Kibble and build new. The old code is hard to read, > not to mention to change. I tried it a few times. Good point. What I am trying to uncover here is about setting a common understanding and expectation about what we will and wont do. We need to send a consistent message to people that either uncover problems or bugs to Kibble-1 while the new Kibble is under construction. Do we allow people to report their issue and then use that list as a cross check to make sure that the issue is eitherno longer relevant or has been fixed in the new Kibble? > > > 2) What can we do we ensure that the Kibble dependency in > > reporter.apache.org tool continues to be supported? > > I think once we have a new working Kibble (together with tests > ensuring data quality) we should integrate it into reporter tool - how > I don't know yet. Until that time we reporter should be ok with old > Kibble. This sounds reasonable. The reporter tool is working well as it is and we are only getting comments around clarifications on how the metrics are calculated. > > > 3) How do we plan for the development of the new Kibble? > > Not sure but I would say we should just do it. Once we have some base > on which multiple people can work then we can split the work and so > on. There was no decision if we want to use REST or Graph QL so the > API part is still undefined (thuse I put it out after 1 milestone in > my proposal). But we seemed to be aligned to use Elasticsearch for > database, so the first step is definitely to define some basic classes > and interfaces around it. Just doing it has always been the way :-). I don't the big picture yet and was thinking of putting together something high level that is easy to follow so the information is open and clear for anyone who wants to contribute. Maybe it is as simple as creating the Github Kanban board and a Wiki page link... > > > 4) What can we do to grow Kibble usage and the community? > > Apart from active community we need well-structured project with > enough documentation and developer tools to make onboarding smooth and > easy. > Another good point and I'd be happy to spend some on the documentation. Thanks Sharan > Thanks, > Tomek > > On Mon, 10 May 2021 at 21:32, Sharan Foga <sha...@apache.org> wrote: > > > > Hi Tomek > > > > I think this is a great start. I was thinking about Kibble from a few other > > perspectives too especially around questions or problems that we might want > > to solve. > > > > 1) How far do we support Kibble-1 while the new Kibble is being developed? > > 2) What can we do we ensure that the Kibble dependency in > > reporter.apache.org tool continues to be supported? > > 3) How do we plan for the development of the new Kibble? (And I think you > > have some good responses for this in your message) > > 4) What can we do to grow Kibble usage and the community? (One of the > > things I'd like to include here is CHAOSS and getting Kibble setup as an > > implementation of some their metrics...) > > > > My understanding is that first three each have different repos so perhaps > > we could think about managing the roadmap at that level.. > > > > I will have a bit more time later this week - so let me see if I can refine > > or iterate on this a bit more. > > > > Thanks > > Sharan > > > > > > On 2021/05/02 16:33:04, Tomasz Urbaszek <turbas...@apache.org> wrote: > > > +1 for using Github issues and projects for managing/tracking - it > > > works pretty well in Apache Airflow. > > > > > > Also I'm definitely in favour of defining a roadmap. I think top level > > > points on such a roadmap should correspond to some big milestones we > > > would like to achieve as a community but they also should be small > > > enough to make executing possible. If I may suggest some starting > > > point I would say: > > > bit more > > > 1. Apache Kibble foundations - we have scanners for few data sources > > > and some basic metrics. Data is accessible via underlying database + > > > Apache Superset as dashboard > > > 2. Crucial data sources - more scanners, covering all required data > > > sources (ASF crucial ones). At this point we should have quite good > > > documentation and understanding of the data we are collecting allowing > > > us for better decision making for next steps. > > > 3. Kibble API - we have an extensible API that allows users access the > > > data collected by our scanners. The API can be used for data sources > > > management (for example for configuring credentials). > > > 4. Kibble UI - there's an web app that can be used for accessing the data. > > > > > > What do you think? > > > > > > Cheers, > > > Tomek > > > > > > On Sat, 10 Apr 2021 at 18:48, Sharan Foga <sha...@apache.org> wrote: > > > > > > > > Hi All > > > > > > > > In one of the previous threads [1] the suggestion was made to look at > > > > setting up a kanban board on Github to track the pending tasks. Are > > > > people happy to get one of these setup? > > > > > > > > In addition, I'm thinking whether it might be good to create a high > > > > level roadmap and make it visible so it is clear and we can all see > > > > where we want to achieve. It could also help explaining to newcomers > > > > about what is happening with Kibble and what our next milestone will be. > > > > > > > > What do people think? > > > > > > > > Thanks > > > > Sharan > > > > > > > > [1] https://s.apache.org/xzk6s > > > >
