Release 1.13 timeline and schedule

[AishSundar]

Here's a straw-man PR for the 1.13 release schedule. In general the timeline across the calendar quarter for this release is highly crunched due to 2 Kubecons (Shangai, Nov 13-15 and Seattle, Dec 11-13), US Thanksgiving (Nov, 22-23) and the holiday season during last 2 weeks of December.

I have highlighted some main points below in an attempt to explain my rationale around the schedule and welcome any input/feedback:

• The release cycle length is only 10 weeks long (reasons stated above).
• Total feature planning and development time are reduced by a week each (both are in keeping with release 1.9, which was also a shorter cycle).
  • Start of release cycle to Feature freeze is 3 weeks vs traditional 4 weeks.
  • Time between Feature freeze and Code freeze is 3.5 weeks vs traditional 4.5 weeks.
• Code slush and Code freeze will continue to follow a shortened timeframe, of 1 week and 2 weeks respectively, in keeping with 1.11 and 1.12 cycles. But if CI signal is not green as we get toward the intended slush and freeze we may need to elongate them by starting them earlier. This would be communicated early and broadly if it is an incoming risk to the release stability mid-cycle.
• Code thaw is only 3 days long! This is mainly to get the release out before Kubecon Seattle which falls in the 2nd week of December. Releasing post Kubecon (3rd week of December) will be risky due to contributors' potential unavailability and is also the least preferred option in terms of media attention and PR. Hence, aiming for a release in the 1st week of December will still give us some wiggle room to release during Kubecon in the event of possible slip.

With all that said, feature load and planning by SIGs will play the most crucial factor in keeping this release on track. Delivering features on time with quality will help stabilize CI signal faster and release on time. During the feature definition phase we will meet with SIGs to re-emphasize release cycle length and encourage them to gauge and plan their feature aspirations accordingly. We carry on the past intent to have initial docs drafts as well as test cases arriving in conjunction with the code at or ahead of code freeze.

In the final days of the cycle in an attempt to thaw and stabilize the release sooner we may postpone more incoming cherry picks to the point release depending on risk assessments.

[AishSundar]

/hold

[AishSundar]

/cc @kubernetes/sig-release-members

Sig leads:
/cc Jaice Singer DuMars caleb miles

1.12 release team leads
/cc @tpepper @justaugustus @mohammedzee1000 @cjwagner @dougm @guineveresaenger @zparnold @feiskyer @nickchase @kbarnard10

1.13 release team delta vs 1.12
/cc @jberkus @kacole2 @nikopen @tfogo @marpaia @aleksandra-malinowska

[k8s-ci-robot]

@AishSundar: GitHub didn't allow me to request PR reviews from the following users: caleb, Singer, nikopen, DuMars, miles.

Note that only kubernetes members and repo collaborators can review this PR, and authors cannot review their own PRs.

In response to this:

/cc @kubernetes/sig-release-members

Sig leads:
/cc Jaice Singer DuMars caleb miles

1.12 release team leads
/cc @tpepper @justaugustus @mohammedzee1000 @cjwagner @dougm @guineveresaenger @zparnold @feiskyer @nickchase @kbarnard10

1.13 release team delta vs 1.12
/cc @jberkus @kacole2 @nikopen @tfogo @marpaia @aleksandra-malinowska

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[AishSundar]

/milestone v1.13

[nickchase]

Maybe I'm nuts for suggesting this, but because it's such a short cycle, and because we will have ~40 new features in 1.12, maybe 1.13 should just focus on stabilizing existing features...?

[idvoretskyi]

I second @nickchase here. Based on my experience of wrangling the features in Q4'2017, it's the best scenario due to the short release cycle and heavy media attention to the new Kubernetes release during KubeCon.

[AishSundar]

@nickchase @idvoretskyi I completely agree with you. Infact @justaugustus suggesting the same idea of a stabilization release restricting only to features falling out of 1.12. I am not sure who should to run this by for approval before we make that official. @jdumars and @calebamiles do let us know what you think.

[justaugustus]

We won't need this file, as no references to it have been created yet.

[justaugustus]

Let's make sure we bring this to next SIG Release meeting, Aish.

[nikhita]

Should this be Nov 16 since that's the code freeze date mentioned above?

[tpepper]

Not sure what this line is, but I think between create the branch and release the branch, this one's not needed?

[AishSundar]

ah this should have been 1.14.0-alpha.0. Thanks for catching will fix it.

[tpepper]

Could be sparsely attended depending on KubeCon China attendance by release team? I suggest you query ASAP and note privately the release team's travel plans for the two KubeCons (and anything else) so you know ahead of time who's going to be out of hand when.

[AishSundar]

Thanks for the suggestion. Coming to think if it, igf I start weekly meetings from 10/1, I will move to MWF burndown a week earlier i.e., from Nov 5th.

[tpepper]

I'd reckon you're probably ok starting in week 7, if the milestone's work list is kept shorter and less ambitious in the feature definition phase. And might not have too many people at KubeCon China...check and see. Might be a non-issue.

[tpepper]

Mostly LGTM, aside from a couple small things I noted in particular the blog comment and release-1.9 CI comment.

[jberkus]

I'm concerned that 48 hours for code thaw-to-final-cherrypick is about 24 hours too short. We need enough time for folks in multiple geos to review/authorize any trailing cherrypicks or fixes.

[AishSundar]

I agree Josh that we might need a day more but I would like to aim for a shorter thaw and CP period and extend it in the worst case. With the current schedule there is almost a week of buffer(before Kubecon), thus giving us room to extend thaw by a day if need be.

[AishSundar]

@tpepper I have addressed your comments. PTAL

All, I would like to push this draft of the schedule by Thursday 9/27, in time for 1.13 on Monday 10/1. So please provide any feedback by then. Thanks.

[tpepper]

/lgtm

I'm curious to see if any of the broader SIG leads have any worries, but it seems like a reasonable set of compromises given everything on the calendar this last quarter of 2018!

[spiffxp]

I know it's not technically release related, but I wonder if it would be worth explicitly adding some rows in here to remind people when the following are happening:

• Kubecon Shanghai (Nov 13-15)
• US Thanksgiving (Nov 22)
• Contributor summit + Kubecon Seattle (Dec 9-13)

[AishSundar]

done

[jberkus]

/lgtm

[cjwagner]

bump for tide?

[AishSundar]

@spiffxp added the Kubecons and Thanksgiving into the schedule.

@tpepper @jberkus @spiffxp need LGTM again plz

[jberkus]

/lgtm

[spiffxp]

/lgtm

[jdumars]

/lgtm

[tpepper]

/lgtm

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: AishSundar, jberkus, jdumars, spiffxp, tpepper

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [AishSundar,jberkus,jdumars,spiffxp,tpepper]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[AishSundar]

/hold cancel

Thanks all for the review, getting this in now