January 2022 Summaries

Engineering teams often face the challenge of deciding whether to develop tools in-house or outsource them, considering factors such as value and cost. FullStory's team chose to outsource feature management, realizing they could not achieve their desired level of safety for rolling out services with interdependencies on their own. Both FullStory and SPS Commerce implemented LaunchDarkly for feature management, improving their release processes and allowing them to build new features and roll them out in beta to a subset of customers. The adoption process within organizations varies, but excitement and word-of-mouth can help drive the implementation of tools like LaunchDarkly.

Feature flags can enhance software delivery by eliminating dependencies between teams, reducing the scope of changes to be merged and tested, minimizing the number of test environments, and enabling testing in production. They are also valuable for automated testing. LaunchDarkly offers various approaches to using feature flags in testing, including reading flag values from files, wrapping SDKs, mocking the LaunchDarkly client, setting up test conditions with the LaunchDarkly API, and utilizing Relay Proxy in offline mode. Each method requires some effort but can significantly improve the testing process.

Engineering intelligence platform Code Climate's Velocity recommends using code Cycle Time as an objective metric for measuring engineering speed. This measurement focuses on the time between a developer's first commit in a pull request (PR) and when it is ultimately deployed, allowing teams to isolate coding activities from planning and design. By breaking down Cycle Time into four components - Time to Open, Time to First Review, Time to Approve, and Time to Deploy - engineering leaders can identify areas for improvement and work with their team to optimize the software development pipeline. Coaching developers on good coding hygiene, limiting work in progress, and improving communication can help reduce Time to Open and overall Cycle Time. Involving the entire team in identifying bottlenecks and implementing solutions can lead to further optimization and increased engineering speed.

As the new year approaches, Heidi Waterhouse emphasizes the importance of investing in internal technical documentation to enhance team performance and efficiency. Drawing on industry reports such as the "2021 Accelerate State of DevOps" and GitHub's "State of the Octoverse," she highlights that quality documentation can significantly boost productivity, improve security practices, and meet reliability targets. Waterhouse argues that while internal documentation is not something that can be bought, it serves as a crucial tool for reducing task friction, onboarding new employees, and preventing organizational bottlenecks. She suggests that making documentation an integral part of the software development process—similar to testing—and recognizing it in performance reviews can lead to higher-quality outcomes. Waterhouse also shares practical tips for starting or improving documentation efforts, such as conducting a documentation inventory or addressing the most urgent documentation gaps, and stresses the importance of routine maintenance to ensure accuracy. By making documentation a habitual part of workflows, organizations can leverage it as a powerful tool to align team efforts and preserve institutional knowledge.

LaunchDarkly introduces "Teams" functionality to help account admins better manage their organization's use of the platform. Teams enable grouping members into teams, allowing for easier management of member access and permissions. A new role, team maintainer, is also introduced to empower team leads in managing team composition. This feature is available now to Enterprise plan customers, with more improvements planned such as automatic syncing with identity providers and extending existing features to be utilized by teams.

The Accelerate State of DevOps report by DORA highlights that top-performing software development teams carry out code deployments nearly a thousand times more frequently than the lowest performers, often utilizing trunk-based development. Trunk-based development is a version control management practice where developers work in short-lived branches and merge their work into a shared trunk frequently. This methodology streamlines merging and integration phases, reduces batch size, and creates an iterative communication pattern. It also encourages collective flow state, minimizes complexity, and supports collaborative achievement of a flow state. However, adopting trunk-based development may not be suitable for everyone and comes with its challenges, such as managing monolithic repositories and ensuring codebase integrity. Despite these challenges, the benefits of trunk-based development can lead to higher job satisfaction, lower rates of burnout, and a highly-generative working culture.