Scrum of scrums/2014-10-29

Facilitating: Andrew Otto

Partners + W0Edit

  • #145 Make app zero-rating croutons have per-operator colors: Android merged. Next up: iOS.
  • #149 iOS connection management exception handling: code moved around, need to see if the issue is significant for iOS users at large.
  • #146 Lua integration for Partners Portal: slimmed down patch submitted and merged; Yuri working on a follow up patch with features similar to the mega patch, but with more restrictions in permitted parameter types and so on.
  • #152 mdot/zerodot webroot Accept-Language redirects for zero-rated access: in code review, heads up to Ops.


  • Work on UploadWizard upcoming
  • Lots of metrics coming your way - image load times now being measured in samples
  • Structured data also on the move - we may have an early draft of an API as soon as mid-November
  • #150 about Swift logging...we keep dropping the ball on pushing this, but maybe someone from ops (Faidon? Andrew?) can speak as to what's up with dat


  • #148 - Aaron is review, some updates in progress
  • #146 - Zero merged the patch
  • Bryan's email about logstash disks



  • Ongoing bug fixing
  • Parsoid now sends logging events to LogStash
  • Working on request and cpu timeouts to kill cpu hogs and log them


  • Hired Marielle for citoid / citation work
  • Parsoid server in deployment-prep (beta labs) crapped its pants yesterday, Roan had to rebuild a Parsoid node from scratch
    • Dear labs people, please investigate/fix this problem where instances mysteriously die (lose LDAP and become un-SSH-able?)
    • Dear ops people, I took advantage of puppet being broken for Varnish servers in labs, but maybe it shouldn't be?
  • Table editing! Landing in wmf7
  • Roan working on VE instrumentation with input from Ori


  • Hiera coming soon to labs openstack GUI
  • bugzilla and RT migration into phabricator in progress
  • Ganglia moved to a new server
  • elasticsearch downtime on Monday. :(
  • #150 more swift logs - ?
  • #127 elasticsearch hardware - in progress this week!
  • #147 staging cluster? need to ping coren and hashar

Fundraising TechEdit

  • Tweaking France forms
  • Configuring email bounce handling
  • Fixing Central Notice bucketing
  • Deployed Dash preview, fixing bugs
  • Backfilling missing banner impressions


(short list this time)

  • Sprint 77 Ongoing
    • Setting up beta and prod environments for CX deploy

-- prod is just in talks

    • Apertium service/cxserver now working in beta
    • New translation dialog improvements
    • Mark articles published with high MT with category
    • Link editing bugs
    • CLDR updates


  • Working on wikimetrics with Grantmaking
  • data warehouse plans are moving forward, have become a critical part of our plans this quarter
  • Event Logging batching insert work underway


  • Math 2.0 (powered by mathoid) deployed last Thursday, provides MathML render mode [1]
    • Mostly volunteer-driven: Moritz Schubotz (physikerwelt), MathJax team
    • Generally positive feedback, some bug reports
    • Now talking about speech annotations
  • Hiring taking up a lot of time
  • RESTBase has gained generic backend monitoring (errors, response times) via statsd to graphite; GELF logging via bunyan (-> logstash)
    • Not quite ready for deploy yet:
      • need to test & review WIP puppetization, depends on Cassandra puppetization being merged (thanks to Andrew Otto!)
      • some code clean-up that should happen before the deploy
      • hw needs for real production deploy will be based on numbers from testing

Mobile WebEdit

  • Deploying WikiGrok extension
    • Performance review completed
    • Security review later today
    • Working with Dario/analytics on development priorities
  • Still working on templates in core (HTML proof of concept:
  • Wondering about Staging Cluster

Flow, EchoEdit

Nothing new, TOC coming

  • next sprint we are going to develop Flow analytics, will reach out to other teams

Task WallEdit