April 2022 Summaries

LaunchDarkly introduces a feature that allows users to automate percentage rollouts for new features. This approach enables gradual release of the feature to a certain percentage of the user base, reducing risk and allowing time for adjustments. The platform offers two methods for performing percentage rollouts: an automated method using workflows and a manual method for targeting specific groups of users. Automated rollouts can be set up with various parameters such as start date, variation to rollout, percentage increments, and frequency. On the other hand, custom workflows allow users to create tailored release processes, including multiple approvals and user segment targeting. LaunchDarkly plans to release workflow templates in the near future for reusing saved workflows.

Continuous testing is an automated end-to-end testing solution that integrates into existing development processes, eliminating errors and supporting continuous integration and delivery (CI/CD) pipelines. It involves testing early, often, and at all stages of the life cycle with automation. The main goal of continuous testing is to assess business risk coverage and provide instant insights into whether a software release candidate is too risky or good to go. Continuous testing accelerates software delivery by running automated tests every time new code is committed, improving code quality, reducing feedback cycles, and minimizing human error. It also helps in managing the complexity of scaling applications and maintaining continuous feedback loops. Important types of continuous testing include unit testing, integration testing, and regression testing. Production testing with feature flags using tools like LaunchDarkly can provide valuable benefits to organizations as part of a robust continuous testing strategy.

LaunchDarkly, a feature management platform, is hosting its first in-person user conference called Galaxy Tour from June through October 2022 across five cities in the United States. The event aims to improve users' understanding of both LaunchDarkly and feature management processes. It will cover topics such as operationalizing and implementing feature management, leveling up experimentation capabilities, establishing effective workflows, leveraging feature flags, and integrating feature management into everyday processes. Full dates and venues for the cities are provided, with more details to be announced soon.

LaunchDarkly provides SDKs for various languages, frameworks, and platforms that not only connect applications to its API but also offer additional functionality such as streaming flag changes and local caching of flag values. While the SDKs are straightforward to use, developers may prefer creating custom wrappers around the API provided by the SDK for reasons like encapsulating all SDK-specific interaction in a single library, simplifying SDK usage, preventing accidental bugs, or adding domain-specific business logic around flag usage. The example code on GitHub demonstrates how to create such wrappers using Node.js Express project with wrappers for both the frontend browser client and the backend Node server environment.

LaunchDarkly has introduced support for big segments, allowing users to import and export segments containing tens of millions of users via CSV files. These large segments can be used for targeting and experimentation purposes directly from the UI. This feature aims to make user targeting more scalable and efficient. Users can learn more about getting started with big segments in LaunchDarkly's documentation, and feedback on this or other features is welcome at [email protected].

LaunchDarkly has released a Visual Studio Code extension that allows developers to manage feature flags directly within the VS Code interface, eliminating the need for constant switching between code and flag dashboard. To use this extension, users must have a Pro or Enterprise plan on LaunchDarkly. The setup process involves installing the extension, creating an access token in the LaunchDarkly dashboard, configuring the settings in VS Code, and then using the extension to view, manage, and toggle feature flags directly from within the IDE. This extension is expected to save time and improve efficiency for developers who frequently use feature flags during development and testing.

LaunchDarkly is a feature flagging system that aims to help software teams manage scalability during growth phases. The platform offers features such as Teams, Big Segments, and Feature Workflows to enhance user targeting, segmentation, and automation. Teams allows for organization-wide adoption with easy onboarding and access control management. Big Segments enables advanced targeting at scale by supporting massive numbers of targets from external data stores. Feature Workflows offers a visual way to automate feature flag rollouts, allowing teams to build complex workflows and define custom release strategies.

Kubernetes and LaunchDarkly serve different purposes but work together to enhance application delivery. While Kubernetes manages the running state of container workloads, LaunchDarkly controls the software state by integrating within applications to wrap parts of its code in feature flags. This allows for targeted testing and rollout of features without affecting all users. By using both tools, developers can build resilient applications and manage deployments more effectively.

This article provides a comprehensive roadmap for migrating to LaunchDarkly's feature management platform from a homegrown or open-source solution (OSS). It outlines potential challenges such as addressing technical debt, prioritizing migrations across teams, and training team members. The migration process is divided into three phases: planning, executing, and post-migration. Tips for each phase are provided, including gradual migration, manageable timelines, choosing the right team for the pilot, standardizing a naming convention, keeping most legacy flags in the legacy system, adding new flags to LaunchDarkly only, changing the wrapper not the calls, dividing into projects, defining your process, using tags, and asking for help. The article emphasizes the importance of planning before executing the migration to avoid technical debt and ensure a successful transition.

The article discusses the importance of having engineers on call to ensure excellent service delivery and customer satisfaction. It highlights how companies can modify their on-call rotation process as they grow, using LaunchDarkly's experience as an example. The company transitioned from a two-team model to a squad model with service ownership, leading to changes in the on-call responsibilities. The new process involved regular business hours coverage by squad members and off-hours coverage by a volunteer-based group of engineers called Virtual On-Call squad. It emphasizes the need for an engineering culture that fosters learning and psychological safety when onboarding new engineers into the on-call rotation. The author shares their personal experience as a first-time on-call engineer, noting that most days were quiet with few alerts, but it was thrilling and provided opportunities to learn.