[Starlingx-discuss] Community activity dashboard
Hi StarlingXers, I just pushed some improvements to the default dashboard at: https://starlingx.biterg.io/ Instead of showing git commit activity (which tends to introduce significant skew when repositories are forked or reused), it's now tracking development activity using merged Gerrit changes, which is much more accurate (and also a more comparable metric to what we use in OpenStack). The "key metrics" numbers on the top-left can be used in conjunction with the date range selection (on the top right) to extract yearly activity numbers or per-release activity numbers. Finally I added three panels at the bottom that show the monthly evolution in corporate diversity for proposed changes, code reviews and ML posts, which is I think a good way to track our progress there. NB: The large number of "unknown" reviews on the middle bottom graph is a glitch that should disappear soon (Zuul review comments counting as unknown instead of being ignored). Comments welcome ! -- Thierry Carrez (ttx)
This is very cool. Thank you! I'm looking at the Diversity graphs using Chrome and it's hard to tell which company is which. Could you add a legend to those charts? brucej -----Original Message----- From: Thierry Carrez [mailto:thierry@openstack.org] Sent: Tuesday, May 14, 2019 7:46 AM To: starlingx-discuss@lists.starlingx.io Subject: [Starlingx-discuss] Community activity dashboard Hi StarlingXers, I just pushed some improvements to the default dashboard at: https://starlingx.biterg.io/ Instead of showing git commit activity (which tends to introduce significant skew when repositories are forked or reused), it's now tracking development activity using merged Gerrit changes, which is much more accurate (and also a more comparable metric to what we use in OpenStack). The "key metrics" numbers on the top-left can be used in conjunction with the date range selection (on the top right) to extract yearly activity numbers or per-release activity numbers. Finally I added three panels at the bottom that show the monthly evolution in corporate diversity for proposed changes, code reviews and ML posts, which is I think a good way to track our progress there. NB: The large number of "unknown" reviews on the middle bottom graph is a glitch that should disappear soon (Zuul review comments counting as unknown instead of being ignored). Comments welcome ! -- Thierry Carrez (ttx) _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
Jones, Bruce E wrote:
I'm looking at the Diversity graphs using Chrome and it's hard to tell which company is which. Could you add a legend to those charts?
The Kibana UI can be a bit confusing. There should be a "<" button on the right of the graph that lets you toggle the legend. If you mouse over the chart panel you should also have an enlarge icon (two arrows in opposite directions) on the top right of the panel that lets you put that graph full screen. -- Thierry Carrez (ttx)
Hi Thierry, sorry for the late response - would you be interested in coming to next week's Community meeting [1][2] to give us an overview of the changes? I'm guessing that'll help generate more awareness and some comments from the community. Bill... [1] call details: https://wiki.openstack.org/wiki/Starlingx/Meetings#7am_PDT_.2F_1500_UTC_-_Co... [2] meeting start time in various time-zones: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20190605T1400 -----Original Message----- From: Thierry Carrez <thierry@openstack.org> Sent: Tuesday, May 14, 2019 10:46 AM To: starlingx-discuss@lists.starlingx.io Subject: [Starlingx-discuss] Community activity dashboard Hi StarlingXers, I just pushed some improvements to the default dashboard at: https://starlingx.biterg.io/ Instead of showing git commit activity (which tends to introduce significant skew when repositories are forked or reused), it's now tracking development activity using merged Gerrit changes, which is much more accurate (and also a more comparable metric to what we use in OpenStack). The "key metrics" numbers on the top-left can be used in conjunction with the date range selection (on the top right) to extract yearly activity numbers or per-release activity numbers. Finally I added three panels at the bottom that show the monthly evolution in corporate diversity for proposed changes, code reviews and ML posts, which is I think a good way to track our progress there. NB: The large number of "unknown" reviews on the middle bottom graph is a glitch that should disappear soon (Zuul review comments counting as unknown instead of being ignored). Comments welcome ! -- Thierry Carrez (ttx) _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
Zvonar, Bill wrote:
Hi Thierry, sorry for the late response - would you be interested in coming to next week's Community meeting [1][2] to give us an overview of the changes?
I'm guessing that'll help generate more awareness and some comments from the community.
Sure, I can make it. -- Thierry Carrez (ttx)
To follow up on the discussion on the call this morning... Here's one example of "Committer" being used rather than author. Jack Ding had helped with upstreaming a lot of our early commits. The biterg page shows him having 120 commits in config, but he's listed as the author of 6: config$ git log --pretty=fuller |grep '^Commit:.*Jack Ding' | wc -l 118 config$ git log --pretty=fuller |grep '^Author:.*Jack Ding' | wc -l 6 -----Original Message----- From: Zvonar, Bill [mailto:Bill.Zvonar@windriver.com] Sent: Thursday, May 30, 2019 11:00 AM To: Thierry Carrez; starlingx-discuss@lists.starlingx.io Subject: Re: [Starlingx-discuss] Community activity dashboard Hi Thierry, sorry for the late response - would you be interested in coming to next week's Community meeting [1][2] to give us an overview of the changes? I'm guessing that'll help generate more awareness and some comments from the community. Bill... [1] call details: https://wiki.openstack.org/wiki/Starlingx/Meetings#7am_PDT_.2F_1500_UTC_-_Co... [2] meeting start time in various time-zones: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20190605T1400 -----Original Message----- From: Thierry Carrez <thierry@openstack.org> Sent: Tuesday, May 14, 2019 10:46 AM To: starlingx-discuss@lists.starlingx.io Subject: [Starlingx-discuss] Community activity dashboard Hi StarlingXers, I just pushed some improvements to the default dashboard at: https://starlingx.biterg.io/ Instead of showing git commit activity (which tends to introduce significant skew when repositories are forked or reused), it's now tracking development activity using merged Gerrit changes, which is much more accurate (and also a more comparable metric to what we use in OpenStack). The "key metrics" numbers on the top-left can be used in conjunction with the date range selection (on the top right) to extract yearly activity numbers or per-release activity numbers. Finally I added three panels at the bottom that show the monthly evolution in corporate diversity for proposed changes, code reviews and ML posts, which is I think a good way to track our progress there. NB: The large number of "unknown" reviews on the middle bottom graph is a glitch that should disappear soon (Zuul review comments counting as unknown instead of being ignored). Comments welcome ! -- Thierry Carrez (ttx) _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
On 2019-06-05 15:08:36 +0000 (+0000), Penney, Don wrote:
To follow up on the discussion on the call this morning...
Here's one example of "Committer" being used rather than author. Jack Ding had helped with upstreaming a lot of our early commits. The biterg page shows him having 120 commits in config, but he's listed as the author of 6:
config$ git log --pretty=fuller |grep '^Commit:.*Jack Ding' | wc -l 118 config$ git log --pretty=fuller |grep '^Author:.*Jack Ding' | wc -l 6 [...]
Keep in mind that Gerrit authenticates committers, not authors, and so can only enforce a CLA or the DCO by mapping the committer's identity to a specific Gerrit account. As a result, pushing a change for another author may circumvent legal protections, so this behavior is generally discouraged and should ideally come with some manual confirmation and affirmation the author really has agreed to whatever legal paperwork the project requires whenever an exception is made. As an aside, the Gerrit account to push the first patch set of a new change is set as the "owner" of that change in Gerrit. This is the value we use when generating electoral rolls, sending event discounts, or building lists of contributors for release announcements. There's a good chance this is also what Bitgeria is using under the hood to associate changes with contributors for the report. -- Jeremy Stanley
Penney, Don wrote:
To follow up on the discussion on the call this morning...
Here's one example of "Committer" being used rather than author. Jack Ding had helped with upstreaming a lot of our early commits. The biterg page shows him having 120 commits in config, but he's listed as the author of 6:
config$ git log --pretty=fuller |grep '^Commit:.*Jack Ding' | wc -l 118 config$ git log --pretty=fuller |grep '^Author:.*Jack Ding' | wc -l 6
There are actually *three* different things. - The overview page tracks the *Gerrit change owner*. As Jeremy says, this is what we use to check CLA/DCO for contributions, and what we use for electoral rolls as well. So it is our default way to count "contribution". - What you compare above is data from the *git repository commits*, not the Gerrit changes. Git commits have two concepts: Committer and Author. Bitergia allows to track both (and by default in their Data sources/git dashboard show Authors). This data (Git Committer and Author) is more unreliable data compared to Gerrit Owners as you potentially inherit work done outside you community (in case of upstream merges) and duplicate data (in case of repository forks). So in summary, you can totally track Git Authorship with the Bitergia tooling.... and I can add a panel on the Git data source page that will make that easier. But I would strongly recommend against using Git Authors and for using Gerrit Owners to count basic code contribution, as it is a much more reliable metric that happens to match how we use in license compliance and governance elections. -- Thierry Carrez (ttx)
participants (5)
-
Jeremy Stanley
-
Jones, Bruce E
-
Penney, Don
-
Thierry Carrez
-
Zvonar, Bill