Services/Meetings/2012-04-03
- Time: Tuesday at 9:15 AM PST / 12:15 PM EST / 5:15 PM UTC.
- Place: Mozilla HQ, Sick Bay (Vidyo room "Services")
- Phone (US/Intl): 650 903 0800 x92 Conf: 98616#
- Phone (Toronto): 416 848 3114 x92 Conf: 98616#
- Phone (US): 800 707 2533 (pin 369) Conf: 98616#
Ops
Hardware
AitC & Token Server UserDB stage gear is in and being built.
Hiring
New Guy. Bob. Yay!
BrowserID
Engineering
Sagrada
Metrics (rmiller/vng)
- Finished metlog-py client message filtering implementation
- Slight refactor of metlog-py configuration code
- Much improved documentation of metlog-py configuration, best starting place for new users
- Significant work completed on metlog-pstree (metlog client extension for gathering process data) and metlog-raven (metlog client extension for python exception data)
- Significant work on metlog-pstree / metlog-raven documentation
SyncStorage/AITC (rfkelly)
Done:
- Properly scripted dev deployment; thanks petef
- Squashed various outstanding SycncStorage bugs
Next:
- loadtesting in dev
- maybe a new Sync 1.1 release...?
Queuey (bbangert/hannosch)
- Metlog integrated into Queuey
- Vagrant VM's updated to run logstash properly, and include RPM packaging utils
- Finally built proper queuey RPM's with all deps
Up Next
- Test full flow of messages through logstash into statsd in VM
- Deploy queuey to our staging VM's
- Put together a brown-bag? Slabs?
Token Server (tarek/alexis)
Last week:
- Did some more packaging + enhancements on the tokenserver
- Refactored powerhose
- Fixed a miscomprehension wrt certificates and browserid
- Started to integrate PyBrowserID with django-browserid
Must have this week:
- [tokenserver] Add Key Derivation in the crypto worker
- [tokenserver] Review the config & redeploy
- [tokenserver] add TTL for the cached certs
- [tokenserver] Finish load testing with multiple clients
- add a "loadtesting" mode using a custom certificate
- [phose] packaging
- [circus] rpm integration
- [circus] release circus 0.2
Good to have this week:
- [tokenserver] add proper caching for the certificates (workers).
- [tokenserver] Use the powerhose worker if available for HKDF (in tokenlib)
- [tokenserver] Review the docs
- [phose] docs
- [cornice] Fix the documentation (the index page doesn't output correctly)
- [cornice] Release the next version
Maybe ?
- [phose] test coverage
- [phose] loadtesting
- [phose] code the c++ version of the worker (again)
Sync
Firefox Sync (gps)
- services-central train landed in mozilla-central (https://tbpl.mozilla.org/?rev=cf60d15e8804)
- Identity refactor (watch out for new bugs)
- Send page title in displayURI client command (BECAUSE ANDROID)
- Misc fixes to bookmarks and tabs records to conform with spec
- Lots of Firefox docs updates (like http://docs.services.mozilla.com/sync/storageformat6.html)
- Refactoring common services/sync bits into services/common (for AITC) bug 731494
- Waiting on patch feedback
- Sync + BID security review Monday April 9 at 1 PM MV time.
- ZOMGBID
Android Sync (rnewman/emtwo/et al)
Notifications
Bipostal (jrconlin)
- Chickenwire injection of Bipostal to BrowserID login screen.
- TODO: update wiring to inject the alias as a valid entry for BrowserID, polish.
Client/Server (jbalogh)
Other Projects (rtilder/nalexander)
QA
Testing and Sign-offs (tracy)
- Client train left the station to m-c last week
- Various Android fixes verified.
Sync Server (jbonacci)
- Initial testing of Sync 2.0, AITC 1.0, and TokenServer on local installs has completed.
- Test Planning for the following:
- AITC in Dev, Stage, Prod, including functional and load
- TokenServer in Dev, Stage, Prod, including functional and load
- Future projects
- BiPostal
- Sync 2.0 API
BrowserID (jbonacci)
- Current train
- Train 24: Bug 740272 - QA and deploy BrowserID train-2012.03.28 to production
BrowserID automation (jrgm)
- This has been formally handed off to WebQA for at least Q2.
Deployment Requests
Support
Metrics
Reports and Monitoring (daniel)
Product
Sync
Firefox 12
Firefox 13
Firefox 14
Firefox 15
Beyond
Identity
Security (dchan)
Marketing
Roundtable
Kilimanjaro
Summary of discussion
There is some concern that with the use of trains and the constant flow of features, that there is a lack of "impact" for new releases. This can be compounded by the fact that there may be little coordination between teams when rolling out features. Kilimanjaro is an attempt to focus product and development on key features that can be turned into major, marketable items. Kilimanjaro allows us to voice to ourselves and the community what we're working on and why it matters.
In order to meet that, we need to make sure that the elements of Sagrada (and all other projects) are properly productized. A goal should be that in 6 months time, we can get a fully understood and clearly spec'd project like Apps in the Cloud and be able to provide a set of tools that will allow that project to be viable to an audience of 10K within weeks. (With the ability to scale higher later, once hardware and other ops elements are made available.)
The work we need to do now is to think about our projects as products and make sure that other folks who are unfamiliar with them have very low bars in order to use them effectively.