Keyboard shortcuts

Press or to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

This topic covers strategies for making your application resilient when feature flags are unavailable due to network partitions or outages.

Application Resilience Goals

Design your application to be resilient even if LaunchDarkly or the network is degraded. Your application must:

  1. Start successfully even if the SDK cannot connect during initialization
  2. Function in at most a degraded state when LaunchDarkly's service is unavailable

A degraded state is acceptable when new features are temporarily disabled or optimizations are bypassed. A degraded state is not acceptable when core functionality breaks, errors occur, or user data is corrupted.

Focus Areas

Resiliency can be achieved through SDK implementation and via external infrastructure components such as LD Relay. This section covers both strategies and their tradeoffs. Our general recommendation is to start with SDK implementation and only add LD Relay if you have a specific use case that requires it.

Future Improvements

LaunchDarkly's ongoing work in Feature Delivery v2 (FDv2) focuses on strengthening SDK and service capabilities so customers can achieve true resilience without depending on additional infrastructure components.