April 2023 Summaries

LaunchDarkly has expanded its Experimentation product capabilities with the introduction of an Experimentation Track within their Technology Partner Program. The program aims to support data from other services and improve experimentation practices for customers. Four launch partners, Tealium, Census, Snowplow, and Sprig, have joined the mission to empower teams to track and measure software releases through experimentation. These partnerships cater to a range of use cases that democratize experimentation within organizations and help achieve faster time to business results on releases. The Sprig, Snowplow, and Census integrations are currently available for customers, with Tealium integration expected in the coming weeks. Interested parties can learn more about these partners on LaunchDarkly's website or apply to become a Technology Partner through their partner page.

LaunchDarkly, a feature management platform, supports various SDKs for languages commonly associated with APIs such as Rust, Go, Javascript, and Python. This article discusses different ways to use LaunchDarkly to roll out new APIs and version existing ones using frontend flags. It covers dynamic address values, dynamic request functions, and the concept of "flags behind flags." Additionally, it explores managing API changes from within the API code itself, including dynamically handling query logic and adding functionality through flags. The article emphasizes that LaunchDarkly's SDK support allows feature flags to be integrated into many different types of APIs and applications.

At the Trajectory conference, Software Engineer Sidney Wijngaarde from Chronosphere discussed their use of LaunchDarkly in managing feature releases, tenant configurations, external client control planes, and instance debugging. The team leverages LaunchDarkly to separate business decisions from engineering ones, manage differences between tenants, control processes without building a one-off config management system, and enable verbose debugging without redeploying or logging. They also implemented reliability tweaks such as a client wrapper for targeting flags appropriately in production and persisting feature flag data in a persistent store to ensure accuracy. As a result, Chronosphere now has 150 feature flags in production, improving efficiency for developers.

LaunchDarkly customers John Hollow, Senior Principal Engineer at Optum; Justin Duhatschek, Progressive Delivery Manager at KBX Technology Solutions; and Mark Eisner-Engebos, Director of Partner Strategy at KBX Technology Solutions shared their experiences with the product during a Galaxy session in Chicago. They discussed how LaunchDarkly helped them get value to customers faster, minimize risk, give more control to stakeholders, enable interconnected teams to move quickly and independently, align on best practices, maintain codebases, audit logs, make on-call less stressful, and measure success through scorecards, internal surveys, and assessing whether things are "on fire or not." The panelists also shared their favorite parts of the LaunchDarkly product and provided an analogy for the classic build vs. buy debate.

In this blog post, Ramiro Berrelleza discusses how feature flags and Cloud Development Environments (CDEs) can be used together to enable feedback earlier in the development process. CDEs allow every team member to have their own copy of an application for testing and validation without being limited by local machine capabilities. By integrating feature flags with CDEs, developers can share previews of changes or improvements with a select group of reviewers at any stage of the development process. This approach allows more people to participate in the feedback cycle, including customers who may not be able to read code. Overall, this strategy helps improve software quality and customer satisfaction by allowing for more efficient testing and validation throughout the development lifecycle.

Feature flags are crucial in managing feature releases as they allow for immediate impact when flipped. This immediacy is particularly important during incidents or when using kill switches, as it minimizes potential business impacts and costs. LaunchDarkly's SDKs enable this instantaneous effect on both the client and server side, making it a valuable tool for feature management compared to homegrown solutions. The SDKs work differently depending on whether they are used in client-side or server-side environments, but both provide fast flag evaluations and updates through streaming connections. A demo showcasing how quickly flag changes can adjust audio playing demonstrates the power of this immediacy. This instantaneous impact allows for quick problem resolution and increased confidence when releasing new features.

The text discusses legacy code, monolithic applications, microservices, their benefits and disadvantages. It also provides criteria for evaluating whether to use microservices or monolithic architecture, migration strategies from monolithic systems to microservices, and tips on getting the most out of a microservices architecture. The text emphasizes that while microservices can offer scalability, autonomy, fault tolerance, and quicker onboarding, they also introduce complexity, harder debugging and testing, and latency issues. It suggests considering factors such as whether services will be created or maintained by another team, use a different programming language, need to stay up if other functionality in the system goes down, or need to be scaled independently of other parts of the system before deciding on microservices. The text also provides two common migration strategies: the strangler pattern and domain-driven design. It suggests using continuous integration and continuous delivery processes for automation during software delivery pipeline and advises against starting with a microservices architecture from the outset, especially when working with a legacy or existing application.

The blog post discusses the power of targeting rules and how they can improve customer experience and streamline developers' work in rolling out new features. Targeting rules abstract business logic away from codebases, allowing users to access new integrations or features without redeploying code. Sunny Guduru, a software engineer at LaunchDarkly, explored this topic during the Galaxy event. The post covers three types of targeting: feature flags or toggles, individual targeting, and user targeting rules. It also highlights the use of custom attributes in user targeting rules to enable more precise control over who gets access to features or functionality. Additionally, it introduces custom contexts that allow teams to deliver to various kinds of things simultaneously with greater precision. The post concludes by sharing best practices for using feature flags and emphasizing the importance of proper documentation and naming conventions.