November 2021 Summaries

The COVID-19 pandemic has accelerated the need for businesses to undergo digital transformation efforts, including updating architecture and infrastructure. Modernization is not a one-time endeavor but an ongoing process that requires constant evolution. Companies often avoid large-scale migrations due to potential risks such as data loss, performance issues, and unexpected challenges. However, using feature flags can help mitigate these risks by allowing for strategic rollouts of new systems and gradual shifts from legacy platforms. Feature flags also unlock benefits like progressive rollouts, improved visibility, and limited impact on users during migration processes.

LaunchDarkly, a feature management platform, integrates well with Amazon Web Services (AWS) to enhance developer workflows. The integration is straightforward and can be seen in various AWS services such as Elastic Kubernetes Service (EKS), Elastic Container Service (ECS), Elastic Compute (EC2), and Lambda. LaunchDarkly allows developers to decouple deployment of workloads from the release of features, giving them greater control over feature releases and rollbacks. It also helps in breaking apart deployment and release tasks, allowing for more efficient management of application functionality. Additionally, LaunchDarkly can be used with AWS Kinesis for data analysis and with pipeline and delivery tools like CodeBuild and CodeDeploy to trigger flag configurations.

LaunchDarkly was recently ranked #47 on Forbes' Cloud 100 list of top private cloud companies. The pandemic has accelerated the consumption and reliance on cloud-based software, leading to a significant rise in both public and private cloud markets. As CEO of LaunchDarkly, Edith Harbaugh discusses how the platform is being used for continuous delivery and modernization initiatives. Traditionally, LaunchDarkly was used for separating deployment from release, enabling practices such as dark launches, continuous delivery, trunk-based development, and testing in production. However, more teams are now using it to enable infrastructure changes like migrating to the cloud, refactoring monolithic applications, and swapping legacy middleware components. The platform helps de-risk migrations by allowing teams to progressively route traffic and control which users experience which system. Harbaugh encourages teams to consider feature management as a way to navigate continuous modernization and replace legacy systems in a safe and predictable manner.

The second day of the annual conference, Trajectory, featured keynotes and discussions on feature management at scale. John Kodumal, co-founder and CTO of LaunchDarkly, highlighted four key principles for effective feature management: developer-first approach, risk-free automation, full-feature lifecycle, and uncompromising performance at scale. The day's talks emphasized the importance of feature management in digital transformation efforts across various industries. Key speakers included industry leaders from NCR, Advance Auto Parts, and iFIT, who shared their experiences leveraging LaunchDarkly to manage code at scale, increase release velocity, and improve customer experience during outages.

The third annual Trajectory conference featured a variety of intellectually-stimulating talks on software engineering and development strategies. Topics included the Flow Framework® for measuring digital outcomes, automating team performance measurements, combining Test-Driven Development with feature flagging, Uber's dynamic configuration platform Flipr, and building corporate culture to foster trust and psychological safety. The event provided valuable insights into improving productivity, efficiency, and overall success in software engineering and development.

The Users dashboard in LaunchDarkly has been enhanced to provide more configuration options, making it easier for customers to find specific users based on attributes or recent activity. There has been a significant increase in the usage of this dashboard over the past year as clients have started incorporating it into their daily operations. Looking ahead, further improvements are planned, including new views into user experiences and enhanced targeting capabilities within LaunchDarkly. Feedback is welcome at [email protected].

LaunchDarkly, a feature management platform, has introduced "Flag Delivery at Edge" to improve its flag delivery network's performance and resilience. The new architecture leverages the Content Delivery Network (CDN) layer of their infrastructure, enhancing caching strategy by moving key services of the LaunchDarkly flag database and initialization system to the edge layer for processing. This change has significantly reduced client initialization time from hundreds of milliseconds to around 25ms, even in challenging connectivity scenarios such as remote geography or mobile devices. The Flag Delivery Network ensures fast and consistent feature delivery across various user experiences, with a focus on security aspects.

The State of Feature Management report by LaunchDarkly provides insights into how over 200 customers use the platform to test innovations collaboratively, standardize safe releases, and accelerate app modernization. Key findings include reduced risk and improved team confidence as major benefits of feature management, with 70% of LaunchDarkly customers releasing several times a week or daily compared to 55% of non-customers. The report also highlights the importance of choosing the right feature management solution, with only 2% of LaunchDarkly users being very stressed when releasing new features compared to almost 20% of non-customers.