October 2022 Summaries

LaunchDarkly's Relay Proxy can address common scenarios faced by development teams such as increased load and high contention in Redis metrics, sudden changes in user groups, and urgent production alerts. The process of integrating the Relay Proxy into infrastructure has been simplified with the release of a Helm Chart for Kubernetes-based deployments. This package manager for Kubernetes allows users to define, install, and upgrade complex applications. Deploying the Relay Proxy can help reduce manual configuration efforts and improve efficiency in handling flag updates. The Relay Proxy chart offers various control options such as scaling and all Relay Proxy configuration settings.

Atlassian's Compass is a developer-experience platform designed to help teams manage complexity in microservice architectures. The LaunchDarkly integration for Compass allows users to visualize feature flags and change history in the context of related components, view active feature flags, and see a timeline of flag modifications. This integration aims to consolidate information about microservices into one place, making it easier for developers to manage their distributed systems.

LaunchDarkly has introduced new team management features, including teams as approvers and flag maintainers, streamlined member assignment during account invitations, and enhanced integrations with Terraform, SCIM for Okta, and SAML. These updates aim to improve scalability and efficiency in managing large organizations with multiple projects and members. The platform plans to continue advancing team-based functionality and making account management easier for complex needs.

LaunchDarkly has officially launched a Cloudflare integration that enables flag evaluation at the edge using its Cloudflare Edge SDK. This integration allows for faster flag variation information delivery within the same Content Delivery Network (CDN) request, improving user experience. The Flag Delivery at Edge feature enhances LaunchDarkly's existing flag delivery network by leveraging CDNs more effectively and reducing client initialization time from hundreds of milliseconds to around 25ms. The Cloudflare integration allows for even faster flag evaluation speeds compared to the standard LaunchDarkly Flag Delivery Network, as it enables feature flag evaluations within the edge layer of a user's network instead of LaunchDarkly's network edge. This integration is currently available to all Enterprise customers and requires LaunchDarkly's Cloudflare Edge SDK.

Deployment strategies are crucial to maintaining constant uptime and minimizing downtime during application updates or feature additions. Two popular deployment methods are blue-green and canary deployments, each with its own benefits and drawbacks. Blue-green deployment involves operating two parallel instances of an application or service, allowing for easy rollback if needed. Canary deployment releases new features to a small subset of users before rolling them out to the entire user base, enabling real-time testing and monitoring. Depending on the use case, a hybrid approach combining elements of both methods can be employed. Feature flags play an essential role in staging deployments and maintaining control over large-scale rollouts.

LaunchDarkly, a feature management platform, has released a new Vue SDK that makes it easier for developers to integrate its features with Vue applications. The Vue SDK provides a plugin for seamless integration with Vue's native features and enables real-time updates of feature flags in the app. To use the SDK, users need to install it using npm or yarn, set up their LaunchDarkly account, and configure environment variables. The new SDK also includes composables for tracking the state of the LaunchDarkly client and reacting to changes in feature flags.

LaunchDarkly's annual conference, Trajectory, is set for November 9-10 this year. The two-day virtual event will focus on software modernization and feature management strategies. Industry experts will share their experiences in software delivery, with discussions centered around six core themes: building & shipping software, collaborating across teams, experimentation, platform engineering, release management, and lessons learned. Keynote speakers include author and researcher Gene Kim, LaunchDarkly co-founder and CTO John Kodumal, and LaunchDarkly co-founder and CEO Edith Harbaugh. Attendees can also participate in workshops to enhance their feature management knowledge and engage with the community. Registration is open on the Trajectory website.

Feature flags are a powerful tool for managing the lifecycle of features and mitigating risks during deployment. They allow developers to control which users see certain features, perform A/B testing, and gradually roll out new features to avoid potential issues. LaunchDarkly is an example of a feature flag management platform that enables multivariate flags, user targeting, and continuous delivery through feature flag-driven development. This approach allows for faster feedback from the market, more efficient collaboration between teams, and better control over the release process. Feature flag-driven A/B testing also enables companies to test backend functionality and new features, providing valuable insights into user behavior and preferences.

LaunchDarkly has launched an education site called LaunchDarkly Academy, offering hands-on, interactive courses and resources designed to help users become experts in using the platform for feature flags and control software. The academy offers three types of resources: Self-guided Missions, Live Missions, and Interstellar Certifications. Users can learn at their own pace with self-guided training or attend live sessions via interactive sessions and virtual labs. Additionally, users can earn certifications to showcase their knowledge and experience in LaunchDarkly. The platform aims to make feature releases safe, boring, and unceremonious for developers.

Technical debt is a metaphor for understanding the trade-offs made in software development that have long-term consequences, similar to financial debt. It can be well-structured and paid off reliably like a mortgage, but sometimes organizations unknowingly take on high-interest loans with no plan to pay them off. Technical debt is not inherently bad, as everything about software engineering involves trade-offs. However, it becomes dangerous when it's silent, unacknowledged, and there's no plan to pay it off. Accumulating and failing to service technical debt can lead to increased costs, stifled innovation, and elevated security risks. Technical debt management requires a multi-pronged approach involving leadership, programmers, and development teams working together to prioritize time for debt paydown, document decisions and trade-offs, and establish guidelines and habits to reduce the system's risk of implosion.