April 2020 Summaries

LaunchDarkly has noticed an increase in the use of feature flags by customers during the pandemic to adapt their services quickly to rapidly changing environments. Feature flags allow for targeted delivery of information based on location, easy disabling of non-functioning code, and adjustments to availability and load management. These tools help businesses maintain reliability and resilience while fulfilling promises to stakeholders in a time of uncertainty.

LaunchDarkly is focused on improving their workflow as developers by enhancing feature management. They leverage feature flags for better control over launching and managing features, allowing real-time updates without lengthy deployment cycles. Customer feedback plays a crucial role in their development process, helping them address pain points and improve their products. The upcoming product launch aims to change how development teams operate, with an online event scheduled on May 14th at 9 am Pacific Daylight Time. The event will feature the company's CTO, customer panel, and Product Delivery team, showcasing new features and technical deep-dives. Registration is available on their website.

Progressive Delivery is a modern software development lifecycle that builds upon the core tenets of Continuous Integration and Continuous Delivery (CI/CD). It enables organizations to ship code faster, reduce risk, and continuously improve the customer experience. Key enablers of Progressive Delivery include feature management, dark launches, feature flagging, testing in production, canary launches, blue-green deployments, A/B testing, and more. The origins of Progressive Delivery can be traced back to Microsoft's "Progressive Experimentation" model for feature rollouts. Core tenets include control over which users see which features, timing of deployment versus release, and who is allowed to release features. Deployment rings at Microsoft and "staff ships" and canary deployments at GitHub are examples of how these principles are put into practice. Progressive Delivery also plays a crucial role in enabling DevOps transformation by increasing deployment speeds, system stability, and cross-team collaboration.

Curve, a London-based fintech startup, recently built a peer-to-peer payments feature called "Curve Send." The team wanted to do a phased rollout of the feature rather than a "big-bang" release. They faced challenges in creating reliable user segments due to social dependencies of the feature, which required participation from both sender and recipient. Curve used LaunchDarkly feature flags to execute a successful phased rollout of the peer-to-peer payments feature. The team learned that when rolling out features with complex dependencies, it's crucial to have fine-grained control over users and ensure that both parties involved in the interaction have access to the feature at the right time.

This week's discussion focused on troubleshooting and debugging in software development. The importance of these processes was emphasized as they are not only used when problems arise but also contribute to discovery and growth. Questions were posed about the weirdest bugs encountered, favorite debugging tools/processes, lessons learned from debugging, fundamental beliefs on the debugging process, and advice for those stuck. Tips shared included using simple console.log() statements for quick debugging and utilizing various debugging tools beyond print statements. The importance of people and processes in fostering a healthy debugging environment was also highlighted. Jennifer's thread emphasized training, empowering, and sharing knowledge with the team to create an effective debugging process. The discussion concluded by encouraging participation in future discussions using the hashtag #ToggleTalk.

The author shares her experience of involving her son in her work during the pandemic when schools were closed. She encourages other parents to interview their children about what they're learning about their parents' jobs, as it provides a unique opportunity for kids to see their parents at work and understand more about their daily routines. Some parents have included their kids in their work activities such as presentations or read-alouds. The author also shares interviews with several children and their parents, discussing what they do for work, how the pandemic has affected their daily lives, and new responsibilities or projects they've taken on. Overall, this article highlights the importance of understanding one another's roles during these challenging times and emphasizes empathy and support among families and professionals alike.

In this episode of Test in Production, Yoz Grahame and Heidi Waterhouse discuss the importance of feature flags for testing in production. They explain how feature flags can be used to manage risk during software development and deployment. They also share tips on how to use feature flags effectively, including setting up leader flags and using prerequisites to control access to features based on user context or version compatibility. Additionally, they discuss the challenges of working from home during the COVID-19 pandemic and offer advice on how to stay productive while adjusting to remote work.

The Elastic Stack integration for LaunchDarkly has been launched, allowing users to navigate to the LaunchDarkly dashboard, find annotated flags, and turn them off in less than 200ms. This integration can help improve system stability and customer satisfaction. Additionally, Elastic's flexible search platform enables users to perform analytics on any arbitrary data sent to it, allowing for advanced circuit breakers, impact measurements, and analysis of software deployments. The integration serves as a starting point for more advanced configurations and improved resilience of distributed systems.

The discussion revolved around resilience, defined as the ability to recover or adjust easily to misfortune or change. It was suggested that building resilience requires a sociotechnical perspective, considering not just technology but also people, processes, and organization as a whole. Questions were posed on how to define resilience, build it in systems, increase tolerance for disruption and failure, and derive value from critical events. The conversation touched upon four concepts for resilience: ability to rebound, robustness, extensibility, and adaptability. It was emphasized that humans play a crucial role in sustained adaptability, as they make decisions about building, changing, and adapting systems. Mental models were also discussed as an essential part of resilience, helping individuals understand and interpret relationships between things and update their mental models when encountering obstacles.

In 2020, the following trends are expected to shape software development: Continuous Integration and Continuous Deployment (CI/CD) without continuous release; leveraging existing workflows by integrating with enterprise tools; personalization of user experiences; increased focus on accessibility; and incorporation of scientific thinking into decision-making processes. These trends reflect a growing emphasis on user needs, data-driven decisions, and rapid iteration in the software industry.

Feature flags are a crucial tool for software teams, enabling them to control the deployment of new features with reduced risk and increased efficiency. They allow developers to hide code until it's ready to be activated, facilitating faster iteration and safer releases. This capability is particularly valuable for testing new features with a limited audience before a full-scale release, managing different user experiences, and quickly disabling problematic features without affecting the entire application. Companies like TrueCar, Intuit, and Atlassian have demonstrated significant improvements in deployment frequency, experiment scalability, and customer satisfaction by leveraging feature flags. Despite their advantages, the adoption of feature flags can be inconsistent across companies due to challenges like tight timelines and resource constraints, but platforms like LaunchDarkly are making it easier for teams to implement and manage feature flags effectively, driving substantial ROI through enhanced productivity and customer experience.

Productivity refers to the effectiveness of work done, and it varies among individuals. LaunchDarkly aims to help teams build better software faster, enhancing productivity. The pandemic has caused many people to feel less productive due to adjustments in remote work and personal life. Tips for staying productive include setting mini-goals, creating nested to-do lists, scheduling breaks, and reducing distractions. Self-care can also be considered a form of productivity. Various resources are available to help improve productivity, such as GTD methodology, tools like Trello or Microsoft To Do, and books like "SuperBetter," "The Miracle Morning," "Digital Minimalism," and "Essentialism."

This article discusses the importance of running small experiments to learn new ideas and improve existing features in websites or apps. It highlights that many changes in these platforms come about as a result of experiments, such as Google testing 41 different shades of blue for their links. Experiments lead to new knowledge, discoveries, and changes in behavior, and they help validate our beliefs using metrics and data. The author provides an example of how to run an experiment with JavaScript, focusing on form completions and the use of a progress bar. They emphasize the importance of forming hypotheses, determining variations, segmenting users, running the experiment, and analyzing the results. By incorporating learning through experimentation, developers can improve their understanding of user behavior and make data-driven decisions.

LaunchDarkly, a feature management system, has been used to control and manage the release of features in various applications. The company decided to migrate its public-facing documentation site from a managed service to Gatsby for greater control and a git-based workflow. During this process, they realized that it might be useful to hide certain sections of the site from specific groups of people. They used feature flags to achieve this goal. Feature flags are decision points in code that can change an application's behavior. LaunchDarkly allows developers to ship code to production without revealing it to customers and gradually release features to a growing audience while easily rolling back if needed. The company created a plugin for Gatsby, making it easy to use feature flags in any Gatsby site. The LaunchDarkly plugin for Gatsby is powered by the LaunchDarkly React SDK, which uses server-sent events to stream feature flag updates to a site in real-time. This allows developers to control feature availability instantly without users having to refresh the page. The plugin can be used for various purposes such as gradually rolling out new sections or features of a site, releasing brand redesigns across multiple properties and products, targeting specific user groups with new features, and conducting A/B tests. The LaunchDarkly plugin is open-source, and the company encourages contributions from users.

LaunchDarkly's first #ToggleTalk event was held on April 1st and focused on "my first flag." The discussion included questions about the participants' first feature flags, deployment methods, lessons learned, and how their practices have evolved since then. It highlighted that feature flags can be used for invisible behavior like database migrations and logging levels, not just user-facing features or UX components. Participants were encouraged to coordinate with others when deploying flags as they may impact multiple business units. The event also emphasized the importance of continuous learning and improvement in implementing feature flags effectively.

LaunchDarkly has updated its flags dashboard to help teams better manage their applications' feature flags. The new flag dashboard provides greater context by answering key questions about each flag, such as whether it is enabled and which variations it serves. Additionally, the flag overview makes it easier for teams to compare flag status across multiple environments. New flag status indicators have also been introduced to communicate the state of features more effectively. These updates aim to provide effortless control and confidence in managing feature flags.