OpenFeature -- one flag to rule them all
summary
Shows working code: a Spring Boot demo using JBang and an in-memory controller, a JMX MBean for live flips, then Togglz with request-based activation strategies, and an OpenFeature client backed by Flagd running in Docker Compose. Demonstrates multi-service sync, multi-stage rules and a 50% fractional rollout variant, all actually exercised and observable. Warns that real projects hit ~150+ flags, stale flags, combinatorial test explosion, and that providers need persistence and client-side resilience (defaults) if the flag service fails.
AI-written summary · about
chapters (18)
- 0:00Introduction to OpenFeature
- 1:05What are feature flags
- 2:30Why use feature flags
- 6:24Feature flags vs Kubernetes
- 8:30How to start with feature flags
- 10:13Demo: In-memory feature flags
- 12:44Demo: JMX for feature control
- 15:13Demo: Togglz framework
- 17:16Demo: Activation strategies
- 20:25Multi-process feature flags
- 21:08OpenFeature specification
- 21:50Demo: FlagD provider
- 25:45Demo: Multi-environment setup
- 29:35Demo: Fractional feature delivery
- 33:07Recap and enterprise features
- 34:31When to use which solution
- 34:52Warnings for production use
- 36:59Q&A session
related talks
38mFun with Flags: How OpenFeature Solves Your Feature Flag Headaches
16mLeveraging Feature Toggles for your Microservices
43mFeature Flags with OpenFeature: Faster, More Collaborative Development
31mTurn Me On With Cloud-Native Feature Flags!
26mBuild Progressive Delivery and Safety Into Your Apps with Feature Flags
06mLightning Talk: Mocking Feature Flags in Unit Tests
25mType-safe Feature Flagging in OpenFeature: Lessons Learned From Using Feature Flags at Google
32mOpenFeature Update From the Maintainers