May 2016 Summaries

GlueCon is an annual developer conference near Boulder, Colorado that focuses on topics such as Cloud, DevOps, Mobile, APIs, and Big Data. The two-day event features six tracks per day, allowing attendees to choose their own learning path. This year's highlights included talks on successful development teams, consistent API writing, tying DevOps back to Lean and Agile principles, continuous delivery and rollbacks, access control techniques, hybrid clouds, identity management, feature flagging, and more. The conference also provided opportunities for networking and discussions with industry professionals.

LaunchDarkly introduces Projects, allowing customers to manage multiple software projects under one account. Each project has its own unique set of environments and feature flags. By default, all team members have access to every project within the account. Custom roles can be used for assigning team members to different projects. The Default project is created upon account setup, with additional projects added as needed. Multiple development environments can be created within each project, facilitating feature flag management throughout the entire development lifecycle.

The online real estate marketplace sought to modernize its development process by hiring experienced managers from top tech companies. They decided to implement a feature flagging system, initially considering building one from scratch but ultimately choosing LaunchDarkly. This allowed the company to switch to trunk-based development and release features with greater precision and control. The team can now roll out features incrementally, monitor their performance, and quickly adjust or disable them if necessary. Overall, this has led to increased flexibility, innovation, and time savings for the company.

In this episode, Edith and Paul discuss the role of programmers in today's world, with a focus on the increasing ease of programming due to tools like AWS and Heroku. They also touch on the left-pad incident and how it relates to microservices. Edith argues that while some may argue that only "real" programmers write the tools, there is value in people who use these tools to build innovative things. Paul agrees, noting that as programming becomes easier, people will be able to build complex applications with minimal coding knowledge. They also discuss how the increasing ease of programming has led to a greater focus on product management and user experience.

The text discusses a method for managing and cleaning up feature flags after rolling out new code to all users. It involves writing the new feature on a short-lived feature branch, then creating a second branch that removes the feature flag and leaves only the new code. Naming conventions are suggested for clarity. After submitting pull requests for both branches, the cleanup branch is merged when it's safe to do so based on flag status indicators. This approach ensures that the work of removing the flag is done while all context is fresh in mind, making it easier to ensure nothing is forgotten.

LaunchDarkly, a feature flag management company, recently achieved a significant milestone by serving billions of feature flag events daily for businesses worldwide. The platform helps companies improve DevOps by managing code deployment separately from business logic and allowing control over who receives specific features at the right time. LaunchDarkly supports various languages and platforms, enabling organizations to unite product management, engineering, QA, and operations. The company has participated in several events like Microsoft Build and plans to announce new features soon.

DevOps is a movement that aims to bridge the gap between development and operations teams by promoting collaboration and communication throughout the software development lifecycle. It was first introduced in 2008 by Patrick Debois, who sought to resolve conflicts between developers and system admins. The philosophy of DevOps centers around continuous delivery and deployment, with a focus on building, testing, and releasing software incrementally. DevOps 1.0 established the initial identity of the movement, emphasizing the importance of trust, cross-functional teams, DVCS, and robust QA/continuous delivery systems. DevOps 2.0 extends these benefits to non-technical team members such as marketing, sales, and product teams by decoupling feature rollout from code deployment. This allows for more targeted user testing, beta programs, and A/B testing of functionality, among other benefits. DevOps tools have emerged to facilitate coordination between previously disparate disciplines, enabling organizations to deliver better software faster while mitigating risk. The future of DevOps 2.0 may see the development of specialized tools that coordinate the skillsets of both developers and non-developers, further enhancing user-centered deployments in alignment with increasing consumer expectations for speedy feature updates and bug fixes.

The Continuous Lifecycle London conference was a sold-out event focused on continuous delivery and DevOps. It featured informative talks by industry experts, including keynotes by Jez Humble and Katherine Daniels. One of the key topics discussed was effective feature flagging to speed up and de-risk development. The pitfalls to avoid when using feature flags were also highlighted. Another recurrent theme in the conference was "Teams, not Tools," emphasizing the importance of communication and collaboration between developers, operations, and customers. The event was well-organized with engaging content suitable for anyone interested in improving organizational efficiency and customer value delivery.

Database migration can be a daunting task for developers, but feature flags can make the process easier and safer. By using feature flags to control reading and writing to each database independently, developers can test their new data store with real, live data while maintaining the integrity of the old store. This approach allows for gradual rollout and monitoring of performance metrics and error logs, ensuring a smooth transition from one database to another. LaunchDarkly successfully implemented this technique during a recent migration from MongoDB to DynamoDB, demonstrating its effectiveness in real-world scenarios.