March 2022 Summaries

Searching for feature flags across different environments within a project is beneficial as it provides an overview of flag targeting states, helps during development by identifying existing flags, and assists product management in tracking features turned on. There are multiple methods to search for flags, including the default method via environment switching, quick search functionality, configurable filters, and API searches. Each method offers unique advantages depending on the user's specific needs and use cases.

As companies scale, their needs change, and they may outgrow in-house tools like feature flag management systems. Key indicators of needing a commercial solution include support for multiple programming languages, client-side vs. server-side flag evaluations, access control and permissions, advanced targeting rules, audit logs, integrations with existing tools, and an intuitive user interface. Additionally, scaling the infrastructure to support feature flagging on a larger scale can be costly. Companies should consider whether maintaining an in-house solution is worth their time and resources, as commercial solutions may offer more comprehensive features and allow teams to focus on core product development.

LaunchDarkly celebrated Black History Month with a series of virtual events centered around "Black Health and Wellness" and the expression of "Black Joy." The company hosted a viewing of the documentary "The Gospel According to André," held a Black Food Fest, supported local Black-owned food businesses, and created a fund for non-profit organizations working on mental health and wellbeing of Black people. Additionally, they hosted an event discussing intersectionality in everyday life with external speaker Madison Butler. The company aims to create an inclusive environment where employees can feel happy and supported.

This article discusses the pros and cons of using a homegrown, open-source, or hosted feature management platform (FMP). It emphasizes that while building your own FMP might seem simple, it's often more economical to use an existing open-source product. The author walks through scenarios of what implementing your own FMP might look like and highlights the importance of feature flags in separating deployment from release. They also debunk the common misconception that feature flags are just configuration settings. The article concludes by recommending using an existing open-source alternative over building one from scratch, as it's usually less economical to run your own flag system.

Monoliths can be a good starting point due to their simplicity and resource efficiency, but they struggle with scalability as an organization grows. Microservices are becoming more popular and offer benefits such as faster deployment, increased agility, and improved resilience. However, transitioning from monolithic architecture to microservices requires careful planning and consideration of factors like time constraints, team size, uptime requirements, and the existing system. It's crucial to involve all stakeholders in the process and ensure that everyone understands the trade-offs and desired outcomes.

LaunchDarkly's engineers, Mike Zorn and Justin Caballero, presented at AWS re:Invent 2021 on achieving a seamless database migration with zero downtime using feature flags. Their approach involved converting a streaming event architecture and migrating production databases gradually, reversibly, and verifiably without affecting user experience. Faced with the challenges of database migrations, LaunchDarkly aimed to improve disaster recovery by transitioning from Mongo and PostgreSQL to CockroachDB. To ensure a smooth transition, they employed a dual-mode operation allowing the application to interact with both old and new databases simultaneously, with the new database initially in "shadow mode." Feature flags provided immediate control over the migration process, enabling gradual changes and easy reversals if needed. This method allowed for extensive testing and optimization in production without user disruption, ensuring that their customers remained unaware of the migration.

Svelte is a frontend framework that has gained popularity among JavaScript developers due to its unique approach and performance benefits. Unlike React and Vue, which do most of their work in the browser, Svelte pushes this to the build step, reducing the size of the JavaScript bundle and potentially improving application performance. Integrating Svelte with LaunchDarkly is straightforward as it only requires using the existing JavaScript libraries for client-side code or the Node SDK for server-side code in SvelteKit. By leveraging feature flags, developers can easily manage and control features within their applications without having to release new code updates.

Technical debt is a common occurrence in software development and refers to the cost of reworking, refactoring, or rearchitecting decisions made due to speed, budget, or limited options at the time. It can be influenced by factors such as not having processes or workflows in place to identify old feature flags, lack of visibility into the state of features, and shifting priorities. While best practices around naming conventions and workflows can mitigate some challenges related to technical debt from feature flags, it is crucial to consider these decisions early on in the decision-making process. The build vs buy question for feature flagging solutions should also take into account unknown costs, flexibility for change, and user needs to minimize technical debt accumulation.

In this episode, James Governor and Cody De Arco discuss the future of progressive delivery and how it can help businesses make better mistakes tomorrow. They also talk about the importance of focusing on customer needs rather than building feature management platforms in-house. Additionally, they share some examples of companies that have successfully implemented progressive delivery practices.

LaunchDarkly introduces a new feature called the Inbox for streamlining flag change approval processes. This feature allows team members to review flag changes as soon as they open LaunchDarkly, with notifications visible on the bell icon. The Inbox displays pending flag change requests and provides detailed information upon clicking "View Details." The aim is to make change management more efficient, and feedback on this new feature can be sent to [email protected].