Scrum of scrums/2015-04-29

Jump to navigation Jump to search




  • What will your team do before we meet again?
    • Release previous sprint's work (bug fixes & pages issues/disambig)
    • Drop iOS 6 support
      • Remove legacy code and allow us to use newer APIs for handling HTML and doing networking
      • Support Swift (also important for upcoming hackathon)
    • Continue developing link preview prototypes as part of our quarterly goal
  • Are you about to put something in another team’s way?
    • No

Language (Editing)[edit]

  • Excuse: Kartik isn't keeping well today :/
  • More languages deployed. Work on CX continue.
  • CX will use REST API soon.
  • Dependency:
    • Bunch of Apertium packaging backlog (who else can help apart from Alex K (ops)?)

Search Team[edit]


  • WQS checkstyle
  • CirrusSearch highlighter improvements
  • Elastic plugin improvements queued for bulk release (because reindex)
  • WQS/Blazegraph bug fixing
  • WQS documentation
  • CirrusSearch testing improvements
  • Geo search improvements
  • Job requisitions

To do[edit]

  • WQS release testing
  • Investigate Blazegraph bug (journal corruption), possibly KB-deletion related
  • CirrusSearch testing improvements
  • Geo search improvements
  • Iron out hardware requirements




  • Planning with new teams
  • Patches on cluster


  • We had couple Parsoid-related "incidents" last week -- incident reports filed and action items identified.
  • Code cleanup phase now winding down and we are going to start picking up bug reports in earnest once more, although I still have one technical debt / cleanup item that I am working on.
  • Arlo is going to start working on to help with "section" editing in VE.
  • Scott is working to get OCG to use RESTBase so we can then turn off the Parsoid job queue (currently edit jobs are processed twice: once for v1 API by Parsoid directly, once for v2 API by Parsoid on behalf of RESTBase).
  • Marc finished a bunch of additional work on CSS-based customization of Cite ( ) -- would benefit from VE team taking a look at it so we can collectively figure out how to push that past the finish line.
  • Tim is busy in secure-poll land for the upcoming board of trustees election.


  • Improved performance in Event Logging, we should stop seeing hour-long gaps in our data, we will keep you updated
  • Event Logging outage yesterday due to mediawiki-core change. Lego fixed it right away but we still lost 2 hours of data which we can not backfill. Report here:
  • Brandon Black deployed our VCL change that sets the "WMF-Last-Access" cookie which allows us to calculate uniques with minimal intrusion. We have some preliminary numbers already, but are still exploring the new data

Release Engineering[edit]

  • Nothing blocked
  • Browser test dashboard up and running this week
  • Work on staging has been tabled for now
  • Deployment tooling research continues (meeting weekly)
  • Isolated CI work to build base images
  • Is Multimedia task still blocked by RelEng? (doesn't seem so)

Fundraising Tech[edit]

  • Working on campaign-specific mixins for CentralNotice, Astropay integration, payment queue abstraction, and using mustache templating for forms
  • Two new hires start Monday: Casey Dentinger and Dylan Kozlowski
  • No blockers, but Sherah Smith would probably be interested in people's experiences with localization and mustache templates. (Matt from Collaboration and Jon from Mobile offered to help) Thanks!

Collaboration Team[edit]

  • We've done some work on RTL support.

Technical Operations[edit]

  • FY2014-15 ending & budget running out; make sure to flag upcoming urgent expansions
  • Analytics 15' downtime sometime next week (to be scheduled)
  • - Better backup coverage for X1 database cluster

Reading Infrastructure[edit]

  • Logstash infrastructure upgrades underway (3 new hosts for the Elasticsearch backend)