Platform/2015-Q1-Goals

Platform

General Goals (by Team)

GFX

Layout

Media

DOM

JS

Accessibility

Perf

Measure E10s performance:

• Complete a report/recommendation​ for E10S riding the release trains based on performance and quality metrics
  • Evaluate E10S performance with Talos tests
  • Analyze BHR & other Telemetry data to determine impact of E10S on Firefox & add-on performance
  • Measure and report per-compartment memory+CPU usage to Telemetry. Report top-offenders found in E10S
    • Deliverable at risk, Yoric has had to take sick leave
  • Develop a general Firefox "perceived responsiveness" metric, use it to study E10S

Flash video performance:

• Make Asynchronous Plugin Initialization feature ready to ride the trains
• Add Telemetry to determine potential performance benefits of using Shumway for Flash ads & videos

"Catastrophic" OOM crashes:

• Report number of OOMs in FHR, to confirm users are indeed leaving because of OOMs
• Confirm and help resolve causes of catastrophic write-combine OOMs
• Check for correlations between OOMs and particular gfx configurations

Telemetry:

• Telemetry backend changes to support new E10S/FHR pings
• Make it convenient for other teams to analyze Telemetry data using Spark

Talos:

• Force a decision on each Talos regression within two weeks of the regression being reported

Other perf deliverables:

• Unblock landing of compacting GC by OS.File
• Enable the background hang reporter on the Beta channel

Networking

Mobile

A*Team

See our Trello board for the complete list.

Highlights:

• Support the conversion of mozmill Firefox update tests to Marionette
• MozReview: add autolander integration with Tryserver
• MozReview: improve Bugzilla integration
• Bugzilla: implement an alternate bug view
• Perfherder: develop a UI inside Treeherder that can be used to inspect Talos performance data
• Treeherder: create the ability to distinguish between Tier 1 and Tier 2 jobs
• Android: develop Android 4.4 emulator images that can be used for test jobs
• Prototype a "big data" project that logs detailed information about tests that can be used to normalize test chunks, identify tests that never fail versus those that do, and be of source of other metrics in the future
• Resolve issues that block us from releasing Marionette 1.0

QA

See our wiki for the complete list

Highlights:

• Automate kick off of update testing, automating entire release process save manual spot-checks
• Begin process for converting automated update and l10n tests to marionette in preparation for e10s (all P1 tests converted)
• Help relieve user headaches from Flash powered content
• Bring quality teams closer to engineering and increase technical acumen in high priority areas
• Create set of risk factor metrics to be used across codebase to identify potential areas of relative risk
• Make security testing automation more robust and more streamlined
• Complete payments automation coverage
• Create a community taskforce to do QA on a specific pilot project

Web Engineering

crash-stats

• Remove all dependencies to the `reports` table from Socorro's middleware
• Upgrade Elasticsearch database to version 1.x in AWS
• Prototype new CI / deploy pipeline
• Create workflow for writing support classifiers over crashes
• Add processor rules as requested to support FF diagnosis
• Shrink Postgres DB to < 1 TB
• sync debug symbols to s3
• create EC2 AMIs for Socorro
• add Amazon SQS support for Socorro

DXR

• Stage elasticsearch-based DXR.
• Render files at request time.
• Index trees independently and in parallel. This will enable us to EOL MXR by no longer having one failing tree scuttle the others and by using multiple build boxes to reindex dozens of trees at a reasonable frequency.

SUMO

• Finish BuddyUp Phase 1
• Finish Questions Forum redesign
• Contributor Metrics Dashboard
• Pay down Technical Debt
  • Upgrade to the latest Django (1.7, most likely)
  • Get SUMO on Python 2.7
  • Switch to py.test
  • Stop using test_utils

Input

• Upgrade to Django 1.7
• Upgrade other libraries to appropriate versions now that Input is on Python 2.7
• 12-factor app work (12 Factor App)
• Prepare for AWS move (set up a stage environment in AWS?)
• 1-5 gradient ratings (Gradient Sentiment)
• Thank you page overhaul (Thank you page) (via OPW intern)

Release Engineering / Release Operations / Developer Services

• Virtualize Windows builders and get builds running in the cloud
• Drive some of our builds out of taskcluster
• Schedule reduced test load (depends on A-team)
• Ship production quality 1.0 of MozReview and Autoland
• Scope and begin implementation of release automation changes with the goal to cut release time in half.
• Support OS X 10.10 builds with minimum additional cost

Security & Privacy

see Security Engineering 2015, Q1 Goals

Firefox and Platform Security

Games Program

Program/Project Management