E10s/Status/March25

e10s Weekly Update: March 25

Executive Summary

• Our current Plan of record to enable e10s as of Firefox . This initial GA release will be a cohort release targeting 100% of users without Add-Ons, users without a11y features enabled, and, the non-RTL language versions of Firefox. Talk about release criteria here.

• The has shifted to burning down M9's [1]. M9 tickets block riding the trains for the above populations for GA. M9 is focused on fixing issues derived from our Telemetry A/B experiments such as top crashes, regressions, and backlog items deemed to be blockers by product.

• We have hid stride with Telemetry data from the Firefox 45 Beta experiment to help us evaluate readiness for ; initial data indicates we are in good shape in terms of jank bug 1223780. Though, we need to continue to confirm this data is valid, especially during the second phase of Beta 45 A/B testing. We also are getting help from platform product to audit our current set of performance criteria and help us get a plan in place to manually validate this data in order to confirm that positive data is equivalent to a positive user experience in the real world.

• With the approach of Firefox 46 merging to Beta on March 7, market readiness becomes a central focus. We're kicking off an e10s Cross-Functional weekly meeting as of Friday, Feb 19.

Why yellow? e10s is designated 'yellow' or 'at risk' because:

Although our release criteria has been set since December, we're finding there are enough questions on part of key teams such as release management and product management on the platform side so we're taking the time to fine-tune to be certain everyone is very confident in the POR. Additionally, although the A/B Experiment logistics and incoming data thus far is better than it has been, it's still really new. We need to be 100% certain it's valid. Additionally, we want to ensure that the improvements we may be seeing in data holds weight with users so we are spinning up efforts to make this happen (more details next week).

Also, we need to flesh our a plan for our add-ons population. Although, accomodating users with add-ons is not a blocker for the initial release, it's important we have a plan because we'd like to have code land in 47, 48+ so we need to be ready. We hare getting closer, efforts on add-on compat have been steady for awhile, but we don't have a final plan of record for the client/UX part of the equation.

Highlights and Accomplishments

We recently landed the following M8's in Fx47. Work is complete to deploy the second phase of the Telemetry A/B Experiments testing. We have some great analysis on incoming stability data from the Beta 45 experiment so far bug 1246180. Check out a summary of the engineering happenings this last week, here.

With the help of the quality team/SV, we have been conducting a backlog scrub of all the [2] issues being tracked for e10s. We've made quick progress to be sure there are not blockers or a combination small things that we would consider a negative trend and block to be found. We also wanted to be sure that we could close as many open items as possible. [3] See the report from SV for details.

Next Steps

As we continue to converge towards an eye for Firefox 46 beta, we'll need to start closely tracking M9's. As previously mentioned, focusing on our add-ons population and also being sure that we validate our measurements with real world proof is important. More details on how that will happen will be available soon. Solidifying that we have consensus and confidence in our roll-out plan is also an important next step and will be a major topic in the cross-functional kicking off next week.

Validation for Stability and Performance

• Phase 2 of the Beta 45 experiment is ready for deployment, see details, here http://pastebin.com/cNKAnW2d.
• We have solid data on stability and are taking action, see more [4] here.
• Jank results are positive thus far though we need to be sure to validate in the real world. See prelim details, [5] here.
• Information about add-ons which are causing jank is also incoming and we need to continue to ensure it's being validated. More on this, soon.

• Upcoming Milestones:
  • Feb 12: Phase 2 starts
  • Feb 25: Testing concludes for 45

Look here for more milestones tracking actions around incoming data/any adjustments we need to make

Memory Usage

• See report on Memory Usage which is mostly positive, here.

Add-Ons + WebExtensions

• Devs are looking for open bugs in webextensions API
• Working to converge SDK issues: https://docs.google.com/document/d/1v0tsxR2OS0eKY2mzDvadrkl-WTjNN-wr9qKU8rpJLa4/edit
• (Carry Over)We have a meeting next week with Product to review Use cases then UX/dev priorities/realities for roll-out selection capabilities for e10s with regards to add-ons (allow/deny lists).
• (Carry Over)We need to make progress on creating a white/black/grey list for incompatible add-ons bug 1233455. The first phase of the Telemetry A/B experiment launching on the 27th will be focused on curating data to help create this black list.
• Source of truth for add-ons for priority and status on compatibility can be found, here.

Automated tests

• Module owners need to be sure their tests are e10s compliant.
• Browser Chrome + and Mochi Test Plan (blocks). Need to get to 90%.
• Need to make sure that M-C and M-A are in lock step
• Goal to get all tests signed-off in M/C by Feb 29th
• More specific status available first week of Feb
• Here is a good test-by-test reference:

https://docs.google.com/spreadsheets/d/10UeyRoiWV2HjkWwAU51HXyXAV7YLi4BjDm55mr5Xv6c/edit#gid=1777180571

A11y

• Wiki to track work is, here.
• Currently focusing on test remoting of a11y calls from main process.
• Good progress is being made on a11y work in general; I will be tracking more closely once we're done with M8.

Release Schedule

Current Plan of Record is to Target 100% of the population who do not have add-ons installed or a11y features enabled for Firefox 46.

Date	Trunk	Aurora	Beta	Release
3-30	40 default (working on m5)	39 off	38 off	37 off
5-11	41 default (working on m6)	40 prompt	39 off	38 off
6-29	42 default (working on m7/m8)	41 prompt	40 off	39 off
8-10	43 default (working on m8)	42 default	41 off	40 off
9-21	44 default (release criteria driven)	43 default	42 off	41 off
11-02	45 default (release criteria driven)	44 default	43 off	42 off
12-14	46 default (release criteria driven)	45 default	44 A/B	43 off
1-25	47 default (release criteria driven)	46 default	45 A/B	44 off
3-07	48 default (release criteria driven)	47 default	46 A/B	45 off
4-18	49 default (release criteria driven	48 default	47 (addons & a11y = off)	46 off

Milestones

We are at the point where we are setting milestones weekly; given the complexity of deliverables, we are using a Trackback Schedule to keep everything organized.