Wikimedia Technology Goals, FY2018–19, Q2 (October - December)

Introduction

The Technology Department has a number of annual goals in support of the Wikimedia Foundation's Annual Plan; this work is detailed in the Annual Plan. Our remaining work falls into four broad areas—foundational, sustaining, supporting our technical community, and supporting the overall community health.

All Technology programs fall under the primary goal of Knowledge as a Service/Foundational Strength - evolve our systems and structures, except for TEC5: Scoring Platform and TEC9: Address Knowledge Gaps which fall under the primary goal of Knowledge Equity - grow new contributors and content.

Purpose of this document

Goals for the Wikimedia Technology department, for the second quarter of fiscal year 2018–19 (October - December 2018). The goal owner in each section is the person responsible for coordinating completion of the section, in partnership with the team(s) and relevant stakeholders.

Goals for the Audiences department are available on their own page

Legend

ETA (Estimated Time of Arrival) fields may use the acronym EOQ (End of Quarter) or EOY (End of Year).

Status fields can use the following templates: To do, In progress, Blocked, N Postponed, N Stalled, Partially done, or Done

Technology Departmental programs

TEC1: Reliability, Performance, and Maintenance

Goal Owners: Mark Bergsma; Ian Marlier; Nuria Ruiz; Bryan Davis

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2 by team:

Analytics

Continue upgrading to Debian Stretch Done
Order and configure hardware for dbstore1002's replacement Done
Add prometheus metrics for varnishkafka task T196066 In progress
Working on a strategy and scripts for updating superset (http://superset.wikimedia.org) task T211706 In progress

SRE

Refresh hardware and perform necessary maintenance - will be Done by end of December

SRE / Traffic

ATS production-ready as a backend cache layer will be Done by end of December
Migrate most standard public TLS certificates to CertCentral issuance Done
Increase Network Capacity Partially done, will be completed in Q3 due to ongoing fundraising efforts

RelEng

Determine the procedure and requirements for an automated MediaWiki branch cut Done

Performance

Train feature developers on the use of performance metrics to detect and address regressions In progress
Deliver high-traffic images as WebP Done
Improve Navigation Timing data, by moving it from Graphite to Prometheus Partially done
Expand mobile testing In progress
Expand outreach and engagement to wider Performance community In progress
Test the effect of MediaWiki commits In progress
Ongoing maintenance of components owned by Performance is always In progress; for this quarter's work, we're Done
Anonymized data publishing is N Stalled and deferred to Q3
Research performance perception in order to identify specific metrics that influence user behavior Done with follow-up in Q3

WMCS

Continue replacing Trusty with Debian Jessie/Stretch in infrastructure layer In progress and will continue in Q3
Communicate Trusty deprecation to Cloud VPS community Done
Develop Trusty deprecation plan for Toolforge Done and communicate that timeline to community Partially done
Track progress towards full removal of Trusty from Cloud VPS In progress, will continue in Q3
Migrate 50% of Cloud VPS projects to the eqiad1 region and its Neutron SDN layer Done, will continue in Q3

TEC2: Modern Event Platform

Goal Owner: Nuria Ruiz

Q2 Goals are Done

Detailed status here.

Wrap-up for Q2:

Development of intake service for events whose transport is JSONSchema/http Done

TEC3: Deployment Pipeline

Goal Owner: Greg Grossmeier

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Formalize the collection of CI infrastructure and tooling metrics is Partially done will continue in Q3 to expose the interface metrics
Develop set of metrics to assess incident reports/post mortems is Done, will probably do follow-up work in Q3
Adopt more services into Deployment pipeline - migrate graphoid to the Deployment pipeline is N Postponed as Graphoid is now recommended for stewardship review, zotero v2 is Done
Deploy blubberoid Done
Reprise the work on the logging infrastructure In progress and will continue in Q3

TEC4: PHP7 Migration

Goal Owners: Mark Bergsma and Ian Marlier

Q2 Goals are Done

Detailed status here.

Wrap-up for Q2:

Ability to serve a % of production traffic from PHP7 are mostly Done with final code reviews In progress
Sampling profiler for PHP7 has been identified and is prepared for use in the WMF production environment Done
Identify and address code issues and opportunities under PHP 7.2 Done

TEC5: Scoring Platform

Goal Owner: Aaron Halfaker

Q2 Goals are Done

Detailed status here.

Wrap-up for Q2:

Upgrade Celery Done
Fix logging for logstash Done
Implement edit quality models for translatewiki Partially done and will wrap up by end of December
Document Feature Injection in The ORES Manual Done
Blog announcement of Undisclosed Paid Editors dataset Done
Resubmit ORES paper to the Journal of Computing N Cancelled
JADE --> Production Partially done to be finished up in Q3

TEC6 Address Infrastructure Gaps

Goal Owner: Mark Bergsma

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Begin the implementation of Q1's Logging Infrastructure design is mostly complete and will be Done by end of December
Expand modern metrics infrastructure coverage is mostly complete and will be Done by end of December
Design and prepare infrastructure for database binary backups is In progress and will continue in Q3
Test Performance implications of MySQL TLS connectivity N Stalled on DBA technology selection/implementation due to other work requirements that have higher priorities
Start migrating watchlist last-view updates is N Stalled due to emergent work and other higher priority work, we hope to get it done in early Q3
Expand Spicerack library and SRE Cookbooks conversion is Partially done and will continue into Q3
Expand Netbox usage Done with stretch goals to be done in Q3

TEC7: Environmental Sustainability

Goal Owner: Erika Bjune

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Identify and contract with an organization that can assess WMF's environmental footprint is Done
Work on an actionable plan for reducing WMF's environmental footprint is now In progress and will be on-going for ~6 months.

TEC8: Search Platform

Goal Owner: Erika Bjune

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Find and hire a contractor to help with NLP work Done
Begin working on one internal NLP project Partially done
Improve autocomplete of Wikidata items Partially done and will continue in Q3
Prototype a feature that is based on collected data Partially done and will continue in Q3
Finish up the Korean morphological library analysis and get ready for deploy into production when ES6 is completed Done
General language support is always In progress
Search for licenses in Commons is N Stalled as we await further instructions from SDoC program
Split the search clusters to increase stability Done
Continue replacing ElasticSearch servers (end of life maintenance) Done
Separate the Wikidata ElasticSearch implementation into a separate extension Partially done and will continue in Q3
Migrate ElasticSearch cluster restart scripts as cookbooks using Spicerack In progress as more testing is needed
Performance and bug fixes for WDQS is always In progress
Service Level Objective (SLO) work for WDQS is In progress and will continue in Q3
Investigate Blazegraph support options and alternatives Done

TEC9: Address Knowledge Gaps

Goal Owner: Leila Zia

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Iterate on and improve the report of the state of the art on bias detection and algorithm audibility is N Postponed
Build a section recommender system based on the section mapping algorithm is Partially done and will be finished by end of December
Build a test API for the section recommendation algorithm is N Postponed until Q3
Improve article recommendation API to completion (of the second stage improvements) is Partially done and will be finished by end of December
Expand the taxonomy of Wikipedia readers is Done
Preparing the infrastructure for conducting the survey is Partially done and will be finished by end of December
Devise the framework for matching newcomers to improve the first design of the framework Partially done and will be continued in Q3
Develop and test a new experiment plan for testing the quality of the algorithm to elicit user interests Done
Finalize the documentation for the research on characterizing Wikipedia readers Done
A series of presentations about the results on characterizing Wikipedia readers Done

TEC10: Build Technical Community

Goal Owner: Bryan Davis

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Plan and visibly improve Toolforge technical documentation is Partially done and will continue in Q3
Survey Wikimedia Foundation staff to gauge interest and support for reviving Tech Talks is Done
Develop plan for Tech Talks reboot is In progress and will continue in Q3
Update visual design and content of MediaWiki.org Main Page will be Partially done by end of January 2019
Support Outreachy Round 17 Done
Support Google Code-In 2018 Done
Review and improve top viewed overview pages of the Action API Done
Submit a proposal for the Wiki Research Workshop Done

TEC11: Support Fundraising Activities

Goal Owner: Erika Bjune

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Support Advancement in all Q2 activities Partially done and will continue in Q3

TEC12: Developer Productivity

Goal Owner: Greg Grossmeier

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

The Annual Developer Productivity Survey results are synthesized and shared, creating a first year baseline is Partially done and will continue in Q3 to get additional feedback

TEC13: Code Health

Goal Owner: Greg Grossmeier

Q2 Goals are In progress

Detailed status here.

Wrap-up for Q2:

Update/refresh review queue is Partially done and will continue in Q3
5 of the 15 prioritized repositories have at least 1 end-to-end test is In progress as the team talks with stakeholders
Assess Platform unit test practices and define improvement plan is In progress and will continue in Q3
Core Platform and Search Platform teams are using TDM PoC In progress and will continue in Q3
Identify key Tech Debt areas and add a process for management In progress and will continue in Q3
Metrics defined and deployed for all 4 Code Health areas Partially done

TEC14: Smart Tools for Better Data

Goal Owner: Nuria Ruiz

Q2 Goals are Partially done

Detailed status here.

Wrap-up for Q2:

Create report for "articles with most contributors" in Wikistats2 To do
Create report for Active editor metrics per project family To do
Provide easier mapping between Wikistats1 metrics and Wikistats2 metrics To do
Provide ability to query metrics per project family in Wikistats2 Done
Add per family unique devices to analytics query service Done
Automatic ingestion from eventlogging data into turnilo datasets that area available for easy exploration Done
Automation of data sanitization for eventlogging schemas in the hadoop backend Done
Presto cluster online and infrastructure accessible by Cloud (labs) users Done
Edit Data Lake Quality - resolve known issues (ongoing) In progress
POC More efficient Bot filtering on pageview data Done
Productionize MediaWiki content processing, ngest and process XML dumps Done

Cross-departmental programs

CDP1: Privacy, Security, and Data Management

Segment 2 - Security

Goal Owner: John Bennett

Q2 Goals are In progress

Detailed status here.

Wrap-up for Q2:

Review and mature our security policies and awareness functions is Done but the phishing campaign is N Stalled to be completed in Q3
Testing campaigns:
- CSP changes are now Done
- 1st round of pen testing (on en wikipedia) is Done
- OIT assessment is N Cancelled, might be picked up in 2019.
- NIST CSF assessment is N Stalled, should be picked up again in early 2019.
- Initial discussion is In progress to include Phan into MW core and should be completed by end of December.
Finalize and test our Incident Response documentation is In progress and will continue in Q3

Segment 3 - Analytics

Goal Owner: NRuiz (WMF)

Wrap-up for Q2:

More restrictive Firewall rules for Kafka. task T204957 N Postponed
Review the requirements for a service implementing a stronger user authentication scheme for the Analytics Hadoop cluster and possibly for other related tools (like Zookeeper). Done
STRETCH GOAL: implement a prototype in labs that the Analytics team can test and evaluate. task T198227 Done

CDP2: Platform Evolution

Segment 7 - Core Platform

Goal Owner: Corey Floyd (WMF)

Q2 Goals are In progress

Detailed status here.

Wrap-up for Q2:

Define and implement a session management service Done

Segment 8 - Core Platform (WMDE)

Goal Owner: Corey Floyd (WMF)

Q2 Goals are In progress

Detailed status here.

Wrap-up for Q2:

Wikimedia Technical Conference: participate and analyze session output is Done and will be published soon

CDP3: Knowledge Integrity

Segment 1 - Research

Goal Owner: Dario Taraborelli

Q2 Goals are < In progress

Detailed status here.

Wrap-up for Q2:

Design a machine learning framework to identify why statements need a citation in English Wikipedia Done
Submit a paper summarizing the modeling work for unsourced statement detection Done
Run the second round of data collection to understand Wikipedia citation usage Done
Prepare the data and analyze the data collected in the second round Done
Perform first round of survey data collection of reader citation usage on English Wikipedia Done
Analyze first round survey data of reader citation usage In progress and will continue in Q3
Host WikiCite 2018 event Done

CDP4: Structured data

Segment 2 - Search Platform

Goal Owner: Erika Bjune

Q2 Goals are In progress

Detailed status here.

Wrap-up for Q2:

Allow search by type of license N Stalled until Q3