"Bitergia to check out the the API of Differential." Is that a commitment from them? If so, what kind of timeline would they be ok with?

In T118753#1843392, @greg wrote:

"Bitergia to check out the the API of Differential." Is that a commitment from them? If so, what kind of timeline would they be ok with?

@Dicortazar, I didn't realize you worked at Bitergia. Do you mind responding to that question ^? Basically, if magic happened and we (WMF) were ready to switch to Differential on March 1st (effectively 2 months from now), would that be OK? (That's not likely, just trying to get an idea :) ).

@Aklapper / @Qgil, I presume this is a blocker to full migration to Differential, from your view?

Bitergia's work is committed through monthly sprints, so all we know today is that this task will not be resolved during DevRel-December-2015. I remember a recent meeting where @Dicortazar or @jgbarah mentioned that they will look into Differential's APIs, but this was said more from a Grimoirse upstream point of view and without any commitment.

@greg, is there an expected timeline for the Gerrit migration? I think it would be useful to align Differential metrics with Gerrit-Migration, but I don't think the potential lack of metrics for Differential should be any blocker for your work. Lack of metrics for Maniphest was definitely not a blocker for the Bugzilla migration last year.

We would like to follow your plans to migrate to Differential as much as possible. During the last monthly meeting, @Aklapper mentioned (if I'm not wrong) that this was very unlikely to happen during the next three or four months. Given this schedule, it is likely that from the upstream point of view, we support this with the new "version" of our tools, which is starting to be published during the next weeks.

[Note: I'm collaborating at Bitergia as well]

According to https://www.mediawiki.org/wiki/Wikimedia_Release_Engineering_Team/Goals/201516Q3, by the end of March the RelEng team will "Garner early adopter projects (goal: 1 project per WMF "team")". In theory these early adopters will be well maintained, so I don't think we need to have metrics support for Differential ready before they jump in. Maybe a reasonable timeline is to have a prototype by the end of March, tested with early adopters and ready by the end of April?

In T114320 at the Developer Summit it was said that there is no exact timeline yet, but "just as a ballpark figure, six months does not sound unreasonable" which would be June/July 2016. However as Quim wrote, earlier testing to iron out some bugs must be possible.

@Aklapper any news about when you guys are migrating to Differential?

@Lcanasdiaz: Not yet. :( I'm waiting for an update in T119908, and it's on @Qgil's list to sort out potential funding related stuff here...

@Lcanasdiaz could you estimate the cost of this project? Let's talk privately.

In T118753#1847441, @Qgil wrote:

@greg, is there an expected timeline for the Gerrit migration? I think it would be useful to align Differential metrics with Gerrit-Migration, but I don't think the potential lack of metrics for Differential should be any blocker for your work. Lack of metrics for Maniphest was definitely not a blocker for the Bugzilla migration last year.

See https://www.mediawiki.org/wiki/Wikimedia_Release_Engineering_Team/Project/Differential_Migration for the up-to-date expected timeline.

And per Quim's comment (which I agree with), I'm removing Gerrit-Migration (and adding Differential).

As discussed in the meeting on 2016-04-08, the first step is Bitergia to mine repositories in Phabricator Differential (which will replace Gerrit) and create and get indexes (Elasticsearch) for Differential.

Followup step, not handled in this task and if we go for Kibana instead of korma:
Displaying/producing panels/widgets. Based on this, Bitergia to produce same indexes and Kibana widgets/panels as for Gerrit in korma (cf.: WMF to decide in T132421). See Also: T132421, T132323, T127078.

In T118753#2099132, @Qgil wrote:

@Lcanasdiaz could you estimate the cost of this project? Let's talk privately.

This is starting to become urgent...

In T118753#2198466, @Qgil wrote:

In T118753#2099132, @Qgil wrote:

@Lcanasdiaz could you estimate the cost of this project? Let's talk privately.

This is starting to become urgent...

Hi guys, sorry for the delay. My partners Jesus & Manrique will send you an estimation asap.

I am tentatively removing this task from the Developer-Advocacy sprints. If you commit to work on it, please bring it back to the corresponding quarter. Thank you.

For Jul 2016 to Jun 2017, we agreed on migrating to the more powerful Kibana based platform (creating panels etc) and to start with the Phabricator Differential (and Maniphest) backends after the first version of the new platform is in place.

Tentatively setting Q4/2016, small chance exists to get this earlier.

Just checking: is this task committed to Developer-Advocacy (Oct-Dec-2016) ?

Thanks for the ping! Very unlikely, hence punting.
It's become less urgent (as RelEng needs more time for Gerrit-to-Differential) and we have a bunch of other issues to iron out first, to make the Kibana version run smooth and allow trusting its data.

I'm trying to get a realistic date for this new panel so I need some feedback from you. What's the status of the Gerrit-to-Differential thing? I saw you guys have around 500 changesets in Differential...

@Lcanasdiaz: Currently WMF's Release Engineering team plans to start migrating a number of projects in January 2017 ("Phase 2 repo migration") and complete the migration until June 2017 ("Phase 3 repo migration").

So I guess that aiming for having basic metrics functionality on https://wikimedia.biterg.io for Wikimedia Phabricator Differential in place around late January or February (if feasible) would be great, and then continue with fine-tuning and sorting out potential issues.

Does that help, or do you need more specific info? If so we could always ask @greg who's already CC'ed on this task.