Patterns & Anti-Patterns for Effective Feature Flagging • Edith Harbaugh • YOW! 2019

Testing

Explore effective feature flagging strategies, covering best practices, anti-patterns, and real-world examples to ensure controlled rollouts, testing, and ultimate success.

Key takeaways

Feature flags are a great way to stop the effects of bugs while fixing the issue.
Feature flags can be used for controlled rollouts, testing in production, and real-world feedback.
Use feature flags as a first-class object, rather than tacking them on as an afterthought.
Feature flags can be used to test different trading parameters, beta testing, and feature soft launching.
Use analytics to determine how many people are using a feature before deciding to sunset it.
Don’t use feature flags for catastrophic errors, as they can make things worse.
Test for ambiguity and conflicting flags, and consider using a kill switch to shut down feature flags if things go wrong.
Think about how you’re going to end a feature, and consider using feature flags to sunset a feature gradually.
Warn people when you’re about to turn off a feature, and consider using a warning period before deprecating a feature.
Consider using feature flags to block access to certain users or features.
Don’t be afraid to turn off a feature if it’s not working, and use analytics to determine if a feature is working or not.
Use feature flags to gradually roll out new features or updates to your users.

Patterns & Anti-Patterns for Effective Feature Flagging • Edith Harbaugh • YOW! 2019

More talks