May 2019 Summaries

Recently, detailed tracing was added to streaming API servers, but the memory required exceeded expectations. To manage this issue, a feature flag was implemented in the instrumentation code, allowing control over whether or not to allocate memory for storing traces. This approach enables on-demand activation of detailed tracing within 200ms and preserves memory when the flag is off. LaunchDarkly's feature flags can be used with custom attributes such as serviceName, operationName, account, project, environment, and key to target specific traces for storage or exclusion. This allows for efficient and customizable tracing in applications without compromising performance.

Designing for failure is crucial when developing distributed systems. One challenge in this regard is ensuring that failure paths are adequately tested. Automated tests can help identify bugs and reduce manual quality assurance efforts. Feature flags can be used to trigger various failure modes, enabling comprehensive testing of all code paths. LaunchDarkly recently implemented a data export service with a resilient retry mechanism using SQS for its event processing pipeline. They also developed an end-to-end method for testing crucial error paths by feature flagging the way events are exported. This approach allows for the simulation of failure scenarios, ensuring that error handling logic is working correctly and providing confidence in the system's resilience to failures.

The text discusses the concept of "The Red Thread," which refers to how humans unconsciously create meaning from their experiences and perceptions. It then delves into key takeaways from LaunchDarkly's first conference, Trajectory. Topics covered include testing in production, experimentation, chaos engineering, processes, and biases. The author emphasizes the importance of having a defined process around testing in production and experimenting to make it less scary. They also highlight the need for fluidity in processes and the removal of biases from them. Lastly, they encourage people to be aware of their biases and take steps to address them when necessary.

LaunchDarkly customers use feature flags for safe deployment of new code and user customization. The company's enhanced Data Export product allows real-time access to full-fidelity feature flag evaluations via Amazon Kinesis, Google PubSub, and Segment soon. This helps in conducting sophisticated analysis for experimentation or A/B testing, troubleshooting bugs, and enriching customer profiles with feature flag evaluations. Data Export connects live streams of LaunchDarkly evaluations to data warehouses like Amazon Redshift, Google BigQuery, Snowflake, etc., allowing quick consumption and analysis of information. Large enterprises can use this for personalization and customized experiences. Data Export is available only to Enterprise plan customers.

LaunchDarkly has promoted its Client-side SDK for React Native to general availability (GA). React Native is a popular framework that enables developers to build Android and iOS apps using JavaScript, similar to how React allows them to create rich web applications. The integration of LaunchDarkly's feature management platform with React Native makes it even more powerful. With the release of version 1.0.0 of the React Native SDK, users can now confidently utilize LaunchDarkly in their apps for handling production traffic. Support for this SDK is on par with existing supported SDKs, and a reference guide is available to help developers get started.