April 2021 Summaries

1-800-FLOWERS.COM, Inc., a leading provider of gifts, has successfully migrated its on-premises infrastructure to the cloud using LaunchDarkly. The company leveraged LaunchDarkly's feature flags to roll out new infrastructure gradually and mitigate potential customer experience risks during the transition. Adopting LaunchDarkly also empowered business stakeholders to control targeting and brand experience, allowing product owners to manage feature releases and run A/B tests. This has led to faster development cycles and reduced risk for 1-800-FLOWERS.COM, Inc., enabling them to continue building their gifting destination while inspiring more human expression, connection, and celebration.

Container orchestration is a process that automates the management of containers, freeing engineers from tasks such as recreating, scaling, and upgrading them. It also helps manage networking and storage capabilities for containers. With the rise in popularity of containers for building software, container orchestration platforms have become essential to reduce operational workload. These platforms can automate various tasks like restarting crashed containers, scaling based on load, and ensuring even deployment across nodes. Managed and self-built (homegrown) options are available for implementing container orchestration platforms. DevOps engineers often focus on these platforms as they abstract infrastructure and allow them to concentrate on one platform. Container orchestration is particularly effective with microservices architecture, which involves splitting applications into smaller pieces that communicate via REST APIs. Docker containers are a popular way of packaging microservices for easy deployment. Kubernetes, Docker Swarm, and OpenShift are some examples of container orchestration platforms. Managed Kubernetes services from cloud providers like Azure, Amazon Web Services, Google, and IBM also exist.

Feature flagging is an essential part of modern software development for controlling code execution and rolling out new features. However, managing feature flags can be challenging due to the need for external context and multitasking between different tools. To address this issue, LaunchDarkly has released a new IntelliJ plugin that allows developers to interact with feature flags directly within their IDE. The plugin provides inline flag information, a toolbar listing all flags in the environment, live updates, metadata refreshes, and right-click options for toggling flag statuses. This helps keep developers focused and efficient while working with feature flags.

Release and deployment management is a system that helps plan, schedule, and control all components of software product launch. It involves decoupling deployment from release using feature flags to enable frequent minor releases, proactive testing in production, and collection of user feedback before a release. The four phases of going live with new software include defining acceptance criteria, dark launching or canary testing, refining the new feature as needed, and finally releasing it to all users. ITIL release management focuses on best practices for ITSM while DevOps release management encourages closer alignment between developers and operations engineers. Key tips to optimize release management process include establishing KPIs, standardizing the release process, reducing risk with feature flags, using automation, and tracking user engagement metrics early and often. Feature flags provide a safety net for continuously delivering software by allowing developers to deploy code whenever they want while enabling businesses to release when ready.

"Unintended Consequences" is a new podcast series created to foster discussions about the complexities and unforeseen effects of scaling systems within the DevOps community, inspired by the lack of personal connections during the past year. Hosted by Kim Harrison, the series aims to bring together experts to share their experiences and insights on managing large-scale systems and the unintended impacts they have on people and technology. Featuring guests such as Charity Majors and Liz Fong-Jones of Honeycomb, Paul Biggar of Circle CI and Dark, and Dr. Aleks Krotoski, a BBC Radio 4 journalist, the podcast explores the surprises and challenges encountered in the evolution of software and sociotechnical systems. With support from Heavybit Industries, episodes are released bi-monthly and include stories of unexpected outcomes and lessons learned, offering a thoughtful look at the intersection of technological growth and human experience.

Test environments are crucial for ensuring that software changes work properly before being released to all users. While testing in production is encouraged at LaunchDarkly using feature flags, some use cases warrant pre-production test environments. This article explores best practices for setting up and navigating these test beds, including building communication into the environment, configuring bug tracing and solution life cycles, leveraging feature flags to test in production, encouraging early testing to shorten error logs, recycling test environments and resources, following a software development life cycle (SDLC) for software testing, and considering testing as a part of deployment. Feature flags can reduce the risk of testing in production dramatically, making it safe to start testing very early in an app's development timeline.

A branching strategy is essential for software development teams working on the same codebase, enabling efficient collaboration while avoiding malfunctions. Common types of branches include trunk, development, feature, release, and hotfix branches. Trunk-based development (TBD) is a popular branching strategy where all developers integrate their changes directly to a shared trunk every day, promoting continuous integration and small changes. Feature flags allow for controlled releases of features in different environments, reducing risk when doing TBD. LaunchDarkly provides feature flag management as a service, enabling teams to use feature flags strategically and seamlessly link them with dynamic releases.

LaunchDarkly recently organized a user conference called Galaxy with the aim of improving communal knowledge and skills around feature management while connecting with each other. The event surpassed expectations, featuring discussions among customers about their use of the platform, sharing of ideas, coaching for fresh voices, and hands-on workshops like Space Camp. The company plans to spotlight various talks in the weeks ahead and encourages those who missed Galaxy to watch the presentations online.

Release management is a crucial process in software development that ensures the optimization of all stages of launching a product or feature, helping teams meet deadlines without compromising quality. The five phases of releasing software include defining specific requirements, specifying acceptance criteria, testing software in production, iterating and refining the product, and finally releasing it to end-users. Key roles in release management include the release manager, product owner, quality manager, and DevOps team. Four best practices for successful launch are standardizing what success looks like, dark launching with feature flags to reduce risk, automating safely, and tracking metrics early and often. Release management tools such as LaunchDarkly can help teams achieve faster releases while maintaining quality.

A test environment is a space where software undergoes experimental uses to ensure its functionality and quality before deployment. It includes the hardware and operating system on which different versions of an application are deployed for testing purposes. Test environments play a crucial role in continuous deployment, allowing developers to manage and maintain applications effectively. There are four main types of environments: development, testing, staging, and production. Each environment serves a specific purpose in the software development lifecycle. Software tests include performance tests, system integration testing, alpha tests, beta tests, unit tests, accessibility testing, user acceptance testing (UAT), and usability testing. These tests help developers identify and fix bugs before deployment. Test Environment Management (TEM) involves setting up and maintaining test environments to ensure their proper functioning and security. Best practices for setting up test environments include focusing on security rather than performance, using feature flags to control access to new features, involving other departments in testing, and establishing effective communication channels among stakeholders. Maintaining test environments requires sanitizing them from each other, using consistent datasets, managing app configurations, optimizing communication, and embracing the philosophy of testing in production with tools like LaunchDarkly's feature management platform.

LaunchDarkly is hosting a free, half-day user conference called Galaxy on Wednesday, April 7th. The event will feature presentations from customers such as RyanAir, Hulu, iPipeline, and Pearson Education, discussing topics like scaling to massive amounts of data and migration best practices. There will also be opportunities for networking with other LaunchDarkly users and sponsors Quantum Metric and Tray.io. Additionally, attendees can participate in a hands-on LaunchDarkly training called Space Camp on Tuesday, April 6th, which costs $100 per person and covers creating feature flags and best practices. This will be the only LaunchDarkly user conference in 2021, so register now to attend.