Testing Kibana: OpenStack code contributions dashboard

We at Bitergia are busy testing new stacks for analyzing and visualizing the software development data we collect. Some our latest tests involve using Kibana for visualization. In this case, we have prepared a dashboard showing the latest contribution data for OpenStack.

Screenshot from 2015-10-19 01-28-46Screenshot from 2015-10-19 01-29-27One of the nice things that these new dashboards allow is the level of filtering and drill down which is possible. For example, in the above dashboard, it is possible to click on any sector on a pie chart, on any entry of a table, on any bar in a bar chart, and the corresponding filter will act. This allows for obtaining specialized dashboards very easily, such as this one with the contributions by RedHat (produced by clicking on RedHat in the list of of top organizations, or the contributions to Liberty, the latest release cycle of OpenStack, by selecting the corresponding period (last bar) in the “OpenStack ten top organizations by release” chart.

If you’re interested in learning about some tips and tricks about what can be done with these dashboards, follow on reading…

One of the most interesting areas in the Kibana real state is the top right, where you can select the dates to be presented.

Screenshot from 2015-10-19 13-38-48Click on it, and you will find a varied menu of timing options, which permits the selection of the period you are interested in studying.

Screenshot from 2015-10-19 13-41-14Another very interesting icon in this top right area is the share button (see below). It allows for sharing the current view of the dashboard as a link. Since the link includes information about the current filters, situation of the charts, etc., you can produce your own version of the dashboard, showing the details of your interest, and then share it so that anyone can have exactly the same view. In addition, you can also bookmark it, and come back any time later to see exactly the same view, maybe with updated data if the underlying database was updated meanwhile.

Screenshot from 2015-10-19 13-55-05This way you can produce for example the dashboard with the activity of a single company in OpenStack (as we showed above for the case of RedHat), or the activity in a certain OpenStack project. As an example, you can link to the activity in Nova, or drilling down a bit more, to the activity of IBM in Nova (Nova is the most active project in OpenStack).

Each person receiving the link can as well change filters and appearance, and redistribute a new link again, which allows for an endless chain of customization and adaption to specific needs.

The dashboard is also completely actionable. You can click on names in tables, in sectors in pies, in bars in barcharts, and the corresponding filters are applied for the whole dashboard, showing almost immediately the new data. The filters applied are always reflected in the top of the dashboard, with some green buttons.

Screenshot from 2015-10-19 14-29-05You can click on those buttons to get extra functionality: deleting the filter, inverting it, or making it permanent (“pinninig” it). The “Actions” menu allows for common actions on all the filters in one operation.

This is not all with respect to managing filters. In the top search box, you can write any term, and that term will act as a filter on all fields in the underlying database. By default, the “search term” is *, which means “anything”, or in other words, search everything in every field.

Screenshot from 2015-10-19 14-33-02You can write a company name, a project name, a developer’s name, or even a term that maybe is in the commit comments, and you will have all the dashboard filtered to include only the commits that include that term anywhere, plus any other field that could be applied, of course. For example, writing “Bitergia”, data is updated to reflect only commits in which Bitergia appears somewhere.

Screenshot from 2015-10-19 14-36-07

In addition to charts and summary tables, such as that for the top companies, Kibana can offer a direct view to the commits being selected in any specific moment. For example, in our dashboard, we have included a table at the end with one row per commit, which allows for a precise view of which data is being represented.

Screenshot from 2015-10-19 14-50-16

Remember that every time you act on the dashboard filtering somehow (selecting dates, using the search box, clicking on actionable areas), all the charts, tables, etc. reflect the new data in a matter of seconds, and that you can share this new view with others, or bookmark it for future use.

Now, go to the dashboard and explore. Remember that this is still only a proof-of-concept of what can be done with the data we’re producing for OpenStack and other projects, when you feed it to ElasticSearch and Kibana. We will have some more news for you while we warm up for the Tokyo Summit.

Meanwhile, you can have some fun paying with these data…

9 thoughts on “Testing Kibana: OpenStack code contributions dashboard

Add yours

  1. Would you please share the procedure how the elasticsearch connect to gerrit to collect data?

  2. Would you please share your way to integrate the kibana with gerrit? I am working on the gerrit report, would like to learn:)

    1. Sanacl, Manrique,

      Thanks for you reply.
      Our company is doing “SVN to GIT transition”, we’d like to collect some data after the svn-to-git cut over and see how many usages. Your tool looks very fancy and it fits our needs. Is it ok we deploy in our company internally?

      I tried some at home, I did catch the architecture of the process. Would you please explain more? Thanks!


      1. You can use it as it is, because it is 100% free, open source software.

        If you want “enterprise” support, just let us know 😉

Leave a Reply

Up ↑

%d bloggers like this: