September 2024 Summaries

The Dev Starter Pack is a bundle of essential developer tools worth over $250k, available for free. It includes ten handpicked tools to help developers build and scale their products across all stages of development. These tools cover authentication, security, communication with users, bug tracking, feature testing, performance monitoring, user engagement tracking, AI integration, database connection, and command line enhancement. The Dev Starter Pack aims to provide a comprehensive set of resources for developers to launch their startups quickly and efficiently.

Despite initial hype, generative AI has recently experienced a period of disillusionment among industry leaders. However, it remains an incredibly powerful tool in many contexts. To solidify the technology's impact at your company and overcome this disillusionment, start by enabling builder teams with generative AI tools that fit closely to their workflows. Be generous in prototyping generative AI features for your own product and double down on automation across all departments. Using automation tools embedded in your systems can create highly relevant and impactful solutions tailored to your business's problems, unlocking new ways of approaching those problems for your team.

This tutorial teaches how to create different user experiences by targeting specific segments within LaunchDarkly using Python. It covers setting up a development environment with necessary tools and dependencies, cloning a repository, installing the SDK key, creating a segment, adding targeting rules, and finally testing the feature flag. The process involves leveraging targeting rules and segments to provide custom user experiences based on various contexts such as email address, plan type, zip code, etc.

Financial institutions are increasingly adopting generative AI (GenAI) applications to improve their operations and customer experiences. However, they must carefully control and optimize these applications due to the sensitive data and regulations in the industry. This article outlines strategies for managing GenAI applications in production, including using feature flags to mitigate risks, iterating on models and prompts with AI-specific feature flags, experimenting to improve performance, and personalizing features for different user segments. By incorporating these strategies, engineering leaders can unlock more potential from generative AI while ensuring safety and compliance.

The text discusses the importance of experimentation in decision making within organizations. It highlights that up to 90% of decisions made without experiments are incorrect and emphasizes the need for more scrutiny on experiment results rather than just relying on intuition or "gut feelings". Experimentation is presented as a way to validate small, incremental improvements and ensure better decision making in the long run. The text also provides suggestions on how to build a culture of experimentation within an organization, such as learning from experiments, building a hypothesis library, and sharing findings with others.

This tutorial teaches how to use LaunchDarkly and Express JS to customize a website based on users' attributes such as .edu email addresses. It covers setting up the environment, creating segments and flags, and using them in an Express application. The process involves initializing the LaunchDarkly client with SDK key, passing user information to LaunchDarkly, and implementing conditional routing for different versions of the website based on user attributes.

This tutorial demonstrates how to add a kill switch feature flag to disable third-party API calls in a FastAPI application using the LaunchDarkly Python SDK. The Dad Jokes API is used as an example data source. FastAPI, a new Python framework, offers benefits such as built-in concurrency support and auto-generated interactive documentation. The tutorial covers setting up a local developer environment with Python and pip installed, creating a requirements.txt file for dependencies, and building the FastAPI application. It also explains how to configure LaunchDarkly kill switch flags and integrate them into the FastAPI application using lifespan events. Finally, it demonstrates how to test the feature flag by enabling and disabling it in the LaunchDarkly dashboard.

The LaunchDarkly CLI now supports a local development server feature that enables developers to create a local copy of their flags for testing in a local environment without an internet connection. This guide explains how to use the local development server with client-side and server-side flags in an Astro application. To set up the feature, users need to install the LaunchDarkly CLI, authenticate it with their account, create a project, and set up two feature flags for testing purposes. The guide also covers how to start the local development server and run the project using it.

A software outage in July cost Fortune 500 companies an estimated $5.4 billion, highlighting the potential devastation of a single bug. To mitigate risks, engineers can employ five strategies: deploy code to production first and release later using feature flags; progressively deliver features to subsets of users; use kill switches as safety nets with feature flags; change broken configurations on the fly with long-term operational flags; and implement monitoring and automatic remediation solutions. These strategies help reduce risks, improve system reliability, and minimize customer impact during software updates or incidents.

In 2024, LaunchDarkly emerged as the leading solution in feature management software, according to G2 Grid® for Feature Management. The platform offers comprehensive feature flag management, advanced analytics and monitoring tools, scalability, flexibility, and strong community support. Its recognition as the best option for both mid-market and enterprise companies highlights its versatility and effectiveness across organizations of various sizes. LaunchDarkly's commitment to user satisfaction and free software offering further solidify its position as an industry leader in feature management.

This tutorial demonstrates how to use LaunchDarkly and Fastify to provide a customized user experience based on the users' email addresses. It covers setting up a development environment with Node.js, installing dependencies, configuring credentials, creating segments and flags in LaunchDarkly, serving static pages from Fastify, and using LaunchDarkly flags in a Fastify application. The tutorial also explains how to target different user segments based on their email addresses and provides additional resources for learning more about LaunchDarkly feature flags.

LaunchDarkly offers a 2-minute Software Release Risk Assessment to help tech companies evaluate the riskiness of their software release processes. The assessment focuses on key areas such as targeting, progressive release maturity, monitoring, and more to understand potential impacts like outages, downtime, poor performance, and subpar user experience. Many organizations underestimate the risks associated with their software releases, making it challenging to identify areas for improvement and implement strategies to mitigate risks effectively. The assessment provides insights into strengths and weaknesses of current release processes, how practices compare to industry standards, specific steps to reduce risk in software releases, and areas where tools like feature flags and progressive rollouts can be leveraged to enhance the release strategy.

Change failure rate (CFR) is a DevOps metric that measures the percentage of changes to production that result in degraded service or require remediation. It serves as an indicator of deployment process reliability and team efficiency in shipping code without causing issues. High change failure rates can lead to financial losses, decreased team morale, customer dissatisfaction, and competitive disadvantages. Common factors contributing to high CFR include lack of proper testing, poor change management processes, insufficient collaboration between teams, complex systems, and absence of gradual rollout strategies. Strategies to reduce CFR involve implementing robust testing practices, adopting continuous integration and continuous delivery (CI/CD), using feature flags for controlled rollouts, improving monitoring and observability, focusing on collaboration and communication, implementing automated rollback mechanisms, and conducting thorough post-mortem analyses. Feature management platforms like LaunchDarkly can help optimize CFR by enabling controlled rollouts, quick rollbacks, A/B testing, environment management, and improved observability.

Mean Time To Restore (MTTR) is an important DevOps metric that measures the average time taken to recover from system failures or outages. A lower MTTR indicates a more resilient system and effective incident response process, while a higher MTTR suggests room for improvement. The article discusses how feature management platforms like LaunchDarkly can help reduce downtime and improve business continuity by enabling quick recovery from incidents and bugs. It also highlights the common misconceptions about MTTR, its relation to other DORA metrics, and various strategies to reduce mean time to restore.

Deployment frequency is a metric that measures how often an organization successfully deploys code to production, reflecting its efficiency and agility in delivering new features, fixes, and updates. It's not just about speed but also creating a sustainable, low-risk approach to software delivery. The deployment frequency spectrum ranges from infrequent large releases to continuous deployment of small changes. Deployment frequency is one of the four key metrics identified by DORA for measuring DevOps performance. Factors affecting deployment frequency include technical infrastructure, automation tools, testing practices, release planning, and cultural factors such as risk tolerance and leadership support. Strategies to increase deployment frequency include implementing CI/CD, breaking down large releases into smaller chunks, adopting feature flags, improving automated testing, implementing IaC, creating a culture of continuous improvement, starting robust monitoring, adopting trunk-based development, and streamlining approval processes. LaunchDarkly's feature management platform enables safe, frequent deployments by decoupling deployment from release, enabling progressive delivery, supporting continuous delivery, and improving rapid incident response.

A recent IDC report highlights how progressive delivery can help engineering organizations safely deliver their code to end users. Progressive delivery involves incrementally exposing new features to internal and external users, allowing for iterations on features and bug detection before affecting a significant portion of the user base. Feature flags enable developers to gradually release features to different user segments, gather feedback, monitor feature impact on specific user groups, and ensure a personalized and optimized experience. Strategies such as percentage rollouts, canary releases, internal testing, entitlement releases, and blue-green deployments are part of a broader progressive release strategy that can be enacted through feature flags and targeting in LaunchDarkly. The report also outlines benefits from de-risking software releases with feature flags and progressive delivery, including more frequent application releases, improved production reliability, enhanced operational resilience, improved confidence across development and operations teams, and better developer and site reliability engineer experience.

The Gartner Hype Cycle for Agile and DevOps, 2024 highlights feature management as a high-value strategy that is steadily gaining adoption. Feature flags are now used not only to disable features but also for experimentation, testing, and progressive release. This allows for continuous monitoring and evaluation of features throughout their life cycle, creating feedback loops that aid in feature design. Additionally, the report mentions the growing use of AI-powered developer tools like GenAI and AI Code Assistants, which can enhance work environments by reducing stress from repetitive tasks and enabling developers to focus on innovative solutions.

In this tutorial, we learn how to use LaunchDarkly's kill switch feature flags in a JavaScript application. We create a new feature for an example fan page and wrap it in a feature flag to mitigate risk during the rollout process. The tutorial also demonstrates how to roll back features if bugs are encountered. By using feature flags, developers can safely add new features to their applications while minimizing potential risks and negative impacts on users.

LaunchDarkly is expanding its platform infrastructure with a new EU region dedicated to supporting data residency requirements for European customers. This move comes after LaunchDarkly joined the EU-US Data Privacy Framework. The company aims to address growing demand for feature management, AI application management, and experimentation in the European market. The new EU instance will leverage hosting infrastructure located within Europe, allowing customers to store their end-user data locally. This initiative is part of LaunchDarkly's commitment to meeting customer needs both geographically and in their feature management journey.