October 2020 Summaries

Code is inherently complex, requiring not only correct functionality but also readability and ease of modification, which often conflict with one another, posing challenges in software engineering. LaunchDarkly advocates for the use of feature flags as a means to mitigate risks and complexities associated with code, though improper use can exacerbate these issues. To aid developers in leveraging feature flags more effectively, LaunchDarkly provides a guide that offers best practices for designing component interfaces, simplifying flag removal, and maintaining code consistency. Their documentation includes a collection of growing best practices and tutorials on creating flags, deployment strategies, and minimizing technical debt, with regular updates to assist developers in optimizing their use of LaunchDarkly in applications.

The text discusses how feature management solutions can empower public sector teams to achieve flexible releases and greater control of user experiences. It highlights examples from Booz Allen Hamilton (BAH) and Ad Hoc, who have implemented feature management for Recreation.gov and Medicare.gov, respectively. Both organizations have reduced the burden on engineering, increased release confidence, and enabled business stakeholders to better respond to user needs while adhering to compliance regulations around security and privacy. The text also emphasizes the importance of being open to change incrementally and exploring options within compliance standards when implementing feature management in the public sector.

1. The "Who Put That Code There?" error - A story about a software bug that caused Knight Capital to lose over $400 million in just 30 minutes. 2. The "Save the Skating Horses" error - A story about how a tiny tweak to Second Life's physics engine led to virtual horses skating past their food and falling off high-altitude platforms. 3. The "Fly Like Superman" bug - A story about a specific kind of heavy bucket in Skyrim that can make you fly when used correctly. 4. The "Taz Dropped the Ball" error - A story about how a software bug caused Taz, an engineer at a large tech company, to accidentally delete all of his team's work. 5. The "I Can See Your Browser History!" bug - A story about how a software bug in Google Chrome allowed users to see each other's browser history. 6. The "My Phone Is Haunted" error - A story about how a software bug caused some Samsung Galaxy S4 phones to randomly take photos and videos without the user's knowledge or consent. 7. The "I Lost My Job Because of a Bug!" error - A story about how a software bug in Microsoft Office led to an employee losing their job. 8. The "My PC Is Possessed" bug - A story about how a software bug caused Windows XP machines to crash with the message "A fatal exception 0E has occurred at 0123:4567 in VXD VMM32 (01)" and then restart automatically, over and over again. 9. The "My PC Is Still Possessed" bug - A story about how a software bug caused Windows XP machines to crash with the message "A fatal exception 0E has occurred at 0123:4567 in VXD VMM32 (01)" and then restart automatically, over and over again. 10. The "My PC Is Still Possessed" bug - A story about how a software bug caused Windows XP machines to crash with the message "A fatal exception 0E has occurred at 0123:4567 in VXD VMM32 (01)" and then restart automatically, over and over again.

In the software delivery lifecycle, a feature flagging platform should have capabilities that cater to different phases such as design, build and testing, release, and launched. During the design phase, broad SDK support and experimentation are crucial for decision-making. For the build and testing phases, granular targeting rules, safety and security measures like role-based access control (RBAC), audit logs, and scheduling release progression are essential. In the release phase, targeted and percentage-based releases, automatic flag disabling upon failure, and filtering relevant flags to reduce technical debt are important features. Finally, for launched features, configuring Slack reminders and finding all flag references in the codebase can help manage technical debt effectively.

LaunchDarkly has introduced Relay Proxy Enterprise, an enhanced version of its Relay Proxy tool that offers better configurability without requiring application restarts. The new features include automatic configuration for detecting changes in environments, projects, and SDK keys, as well as a new in-app UI for managing configurations. These improvements are aimed at making the tool more suitable for large organizations. Upgrades to Relay Proxy version 6.0 or higher will grant access to these new features, which are available to all Enterprise plan users.

LaunchDarkly has introduced a new Azure Event Hubs integration for its Data Export feature. This integration allows teams to stream raw feature flag analytics directly into their databases, enabling detailed customer cohort analysis and monitoring of rollout success with target demographics. The Data Export add-on is available to LaunchDarkly customers on Enterprise plans, which includes access to Azure Event Hubs and four other popular destination types. To get started, contact [email protected] or read the documentation for more information.

In this episode, Tiffany Jacha from Harness joins Alex to discuss continuous integration (CI) and continuous delivery (CD). They talk about the difference between deployment and release, how CI/CD can help with feature development, and the importance of efficiency in software delivery. Tiffany also shares information about Unscripted, an upcoming user conference hosted by Harness. The conversation covers various aspects of CI/CD, including provisioning environments, rollback and release strategies, secrets management, and cost optimization. They discuss how continuous delivery can be achieved through a platform like Harness, which works with any CI platform and supports different cloud providers. Tiffany also talks about the new product, Continuous Efficiency, which helps track spending in the cloud and optimize resource allocation. Towards the end of the episode, Alex asks Tiffany what advice she would give to people starting their CI/CD transformation journey. Tiffany suggests figuring out where the pain points are and focusing on solving those issues first. She also recommends reading the DevOps Handbook and considering value stream management as a way to improve software delivery processes. Overall, this episode provides valuable insights into continuous integration and continuous delivery, their importance in modern software development, and how platforms like Harness can help organizations achieve these goals more efficiently.

LaunchDarkly offers 21 SDKs, including some for browser use like the client-side JavaScript SDK that can be added to websites for feature flag capabilities. While front-end development may pose challenges, adding feature flags to static sites is possible with proper techniques and strategies. An in-depth guide on using feature flags on static sites has been written by LaunchDarkly to address common concerns and provide best practices for optimal website performance. Developers can reach out over Twitter for further assistance or feedback.

LaunchDarkly, a feature management company, has been recognized as #100 on Forbes' 2020 Cloud 100 List. The company is the first to be featured in this category and serves over 6 trillion feature flags daily. LaunchDarkly aims to help customers reduce risk and deliver value through its software solutions. Co-founder Edith Harbaugh expresses gratitude for the recognition and looks forward to moving up the list as they continue to improve their products and services.