KTH/devops-course

A/B testing, feature flags, continuous experimentation

monperrus opened this issue · 24 comments

"Online controlled experiments at large" KDD'2013 about the Microsoft use case

"Holistic Configuration Management at Facebook"

How The Huffington Post uses real-time testing to write better headlines
http://www.niemanlab.org/2009/10/how-the-huffington-post-uses-real-time-testing-to-write-better-headlines/

Dynamic knobs for responsive power-aware computing
https://dspace.mit.edu/openaccess-disseminate/1721.1/72445

About feature toggles:

The modular and feature toggle architectures of Google Chrome
https://link.springer.com/article/10.1007%2Fs10664-018-9639-0

Dark launching has its downsides:

Jane Manchun Wong: App researcher who unlocks new features before they launch
https://www.bbc.co.uk/news/technology-47630849

Bullet Train: Open-source feature flagging
https://bullet-train.io/

The Netflix Recommender System: Algorithms, Business Value, and Innovation
https://dl.acm.org/doi/pdf/10.1145/2843948

There exist different techniques and frameworks to engineer feature flags. They are good topics for a contribution to the course

Unleash: Another Open-source feature flags platform https://www.getunleash.io/

Gitlab includes unleashed. You can controll the feature flags in your project directly from within Gitlab. https://docs.gitlab.com/ee/operations/feature_flags.html

Snowplow Open Source, the world’s largest developer-first engine for collecting behavioral data.
https://github.com/snowplow

The Only Guide to Dark Launching You’ll Ever Need
https://launchdarkly.com/blog/guide-to-dark-launching/

"Split provides everything you need to create, target, and manage feature flags."
https://www.split.io/

A tool for refactoring code related to feature flag APIs
https://github.com/uber/piranha