Scrum of scrums/2016-06-08
Blocked - None
Blocking - Working on External Store on Beta
Expanding our API to fetch notifications by specific pages.
Working on expandable Echo bundles
Split Flow Vagrant role to mitigate some issues with Flow having a lot of MediaWiki-Vagrant dependencies.
Work on ContentTranslation, Compact Language Links continue. Parsoid use of service-runner: working with Marko to deal with final steps (puppet updates, testing in beta cluster).
Followup RFC for Tidy replacement scheduled today (Wed, Jun 8).
Released Parsoid 0.5.1 debian package last week.
Not blocked, not blocking
New repository set up for experimental new extension for file annotations
UploadWizard fixes moving along apace
Gallery slideshow patch looking nicer, proceeding
Reading Infrastructure Edit
Reading Web Edit
Fixing issues with lazy loading images and refs
AB test of Popups extension ("Hovercards") on hu.wiki in progress (50% now)
Mobile Content Service Edit
Deployed first three feed-related endpoints. Not publicly available yet. Use labs instance if needed. Still adjusting response format.
Analytics Edit We're blocking the Release team on its goal to move deployment to scap3, we're trying to prioritize and get it done but it might not fit until early next quarter.
Pageview API is still throwing 500 errors despite the throttling down to 10 requests per second, we've had an update to the AQS compaction (and an update to SSDs from spinning slowpokes) under testing for a while, so far the numbers look about an order of magnitude better.
Ongoing: normalizing and loading all history from all mediawiki dbs into Druid
The team will be at its offsite next week, and Madhu and Dan will be at Wikimania, Andrew taking vacation after the offsite. Our offsite is in Berlin and we're working out of WMDE's offices, so if they need us for any in-person hacking, we're there :)
Technical Operations Edit
ORES is deployed in production. Try it out in
http://ores.wikimedia.org Got a few problems with services on SCB today
A jessie appserver passed smoke tests, we are moving towards upgrading the mw* fleet to jessie now.
GTID 100% deployed on core dbs. Labs/analytics/dbstore needs work.
Insecure POST block still coming up June 12 (this weekend)
service cluster services outage today
approx 15 mins
stil under investigation Cassandra partial outage last week during move to 2.2.6
MathML by default
hit some initial bumps, should be good now
Two usability bugs for
Extension:OATHAuth in progress (T136988, T136989) Two-factor usability survey on hold until after AuthManager is enabled
Fix for T129738 will be deployed this week
Reviews: Use of wheels for deployment by Analytics
Blocking: None as far as we know
Blocked: Reading Infra - AuthManager (blocks 1.27 release) Updates:
Fundraising tech Edit Trying to get payments onto 1.26 before 1.25 EOLs
pondering what to do for the future, including even divorcing DonationInterface from MW entirely Testing new processor in Israel, Japan and Ukraine
Ready to test new PayPal integration method
Enhancing fraud & dos mitigation measures
more work to get off ActiveMQ