Challenges in Distributed Feature Flag Evaluation and How To Solve Them
Talkat OpenFeature Summit Europe 2025 · on Tuesday, 2025-04-01· at 15:20· Level 3 | ICC Capital Suite 2-4·
summary
Examples include a book-printing transaction that must not split jobs, a 4K streaming flag that mustn’t show inconsistent UI, and PII-logging that needs per-request locality. The proposal gives concrete mechanisms: flex set version (monotonic ID propagated via baggage to detect stale rules), propagate rule set or propagate evaluated values to downstream caches for availability, and signed flag payloads (JWKS/JWT) plus SDK validation. Tradeoffs are bandwidth, security and topology exposure; adding flex set version to OpenFeature SDKs is presented as the least invasive first step.
AI-written summary · about
chapters (13)
- 0:00Distributed feature flagging challenges
- 1:54Distributed computing principles
- 4:44Feature flagging inconsistencies
- 6:58Streaming service availability
- 8:02Logging and PII compliance
- 9:46Value vs rule consistency
- 11:12Flex set version for consistency
- 14:19Rule set propagation
- 16:25Context propagation
- 19:15Flex value propagation
- 21:23OpenID Connect for flags
- 22:49OpenFeature’s role in solving these issues
- 25:19Q&A
related talks
16mLeveraging Feature Toggles for your Microservices
53mTinkertoys, microservices, and feature management: How to build for the future
50mFeature Toggle Promotion via CI/CD Pipeline
46mPracticalities of Productionizing Distributed Systems
21mLeveraging Feature Toggles for your Microservices
41mSurviving Continuous Deployment in Distributed Systems
15mStreaming, Fast and Slow: Mitigating Watermark Skew in Large, Stateful Jobs
27mTinkertoys, Microservices, and Feature Management: How to Build for the Future