June 2021 Summaries

Feature flags, also known as feature toggles, are a powerful tool for software development teams, enabling them to conditionally turn sections of code on or off, which facilitates controlled rollouts, experimentation, and rapid feedback. This practice enhances development velocity, reduces risk, and allows for seamless feature management across teams by decoupling feature release from code deployment. Feature flags are beneficial for various organizational roles, including developers, product managers, sales, and support teams, allowing for increased engineering velocity and efficient management of new features. They offer capabilities such as A/B testing, progressive delivery, and the ability to manage betas and early access features. While implementing feature flags can present challenges as they scale, solutions like Harness provide platforms to quickly and effectively integrate feature flags into workflows, offering secure, performant, and scalable management tools that help teams maximize the potential of feature flags throughout the software delivery lifecycle.

Modern open-source frameworks such as Dropwizard and Spring Boot have revolutionized the development of production-ready Java applications by balancing the trade-off between freedom of choice and rapid development speed. Dropwizard is appreciated for its simplicity and community support, while Spring Boot offers extensive integrations within the Spring ecosystem, catering to varied project needs and ecosystem preferences. Both frameworks are highly opinionated, emphasizing convention over configuration to ensure fast application deployment. Dropwizard is known for its use of Jetty for HTTP servers and Logback for logging, while Spring Boot provides more flexibility with embeddable HTTP servers like Tomcat and options for logging frameworks. A key distinction lies in dependency injection support, with Spring Boot offering built-in support, whereas Dropwizard requires community integrations like Guice. Both frameworks provide robust features for metrics, health checks, and task scheduling, essential for production-grade applications. Despite their differences, both frameworks support container-less Java HTTP servers and use fat JARs for simplified deployment. Community support and release cycles differ, with Dropwizard having a larger community and Spring Boot benefiting from structured support and a strong user base. The choice between these frameworks often hinges on the developer's preference for ecosystem integration and specific application requirements.

In an episode of ShipTalk, Santiago Campuzano, a Lead DevOps Engineer at GumGum, discusses his journey from developing distributed Java applications to embracing containerization and Kubernetes, highlighting both the benefits and challenges of this technological evolution. Campuzano reflects on the shift from monolithic architectures to microservices, facilitated by technologies like Apache Mesos and Marathon, and later Kubernetes, for better resource allocation, scalability, and deployment efficiency. He emphasizes the importance of Kubernetes' pluggable architecture, which allows for flexibility but also requires addressing additional challenges such as logging, monitoring, and security. Campuzano notes that while Kubernetes offers significant advantages, many aspects remain under development, leading to a vast ecosystem of solutions that can be overwhelming. He predicts that the Kubernetes ecosystem will eventually stabilize, offering fewer but more mature and reliable options. Throughout the discussion, Campuzano shares insights from his professional experiences, suggesting that patience and skepticism are valuable traits in navigating the rapidly evolving tech landscape.

Angular, a popular front-end framework, can be enhanced by implementing Continuous Deployment (CD) and feature flags to improve user experience and streamline the release process. Continuous Deployment builds on Continuous Integration by automatically deploying successful merges to production, thus enabling shorter feedback cycles and early defect detection. A practical guide is provided on setting up CD for an Angular app using free resources, such as Firebase Hosting for deployment and Travis CI for automation, with a simple CRUD application as an example. Feature flags are introduced as a method to disable or enable features without redeployment, allowing for safer feature rollouts and A/B testing. The guide details how to integrate feature flags using Split's JavaScript SDK, demonstrating their role in supporting a robust CI/CD process by allowing developers to merge unfinished features confidently. Overall, the text emphasizes the benefits of feature flags and CD in fostering efficient, fearless development and deployment practices.

Harness has introduced several product updates to enhance its platform, including the Harness CI Enterprise with Test Intelligence, which optimizes code testing through machine learning, and the Harness Feature Flags, which accelerates feature delivery while minimizing risk. Additionally, Harness has launched Cloud AutoStopping to significantly reduce cloud computing waste by managing idle resources automatically. The updates aim to improve the developer experience by incorporating features such as containerized steps, a visual pipeline builder, and fine-grained role-based access control (RBAC). Furthermore, Harness has expanded its educational offerings with new courses and webinars, and its platform updates have received attention from notable press outlets. The company also announced upcoming features like enhanced deployment capabilities and improved security measures, alongside plans for future events and webinars.

AWS App Runner is a new platform-as-a-service offering from Amazon Web Services that simplifies the deployment of containerized applications by abstracting infrastructure management, allowing developers to concentrate on coding. It integrates seamlessly with Drone for continuous integration and continuous deployment (CI/CD) and Harness for enhanced deployment control, providing a streamlined process from code commit to production. The service resembles Rancher Rio in its ability to create publicly accessible routes for deployments without complex networking configurations. To get started, developers need to set up a Drone installation, an Amazon Elastic Container Registry (ECR) for storing Docker images, and configure AWS App Runner to deploy these images. The integration of Harness with AWS App Runner offers greater flexibility, enabling developers to manage diverse technologies and choose the best AWS services for specific workloads, such as Fargate or EC2 instances, enhancing the deployment experience across various environments.

Harness's second {unscripted} Conference showcased advancements in Developer Experience, FinOps, and Feature Flags, with significant announcements such as Test Intelligence and AutoStopping/AutoStarting capabilities aimed at simplifying and scaling software delivery. The conference emphasized the importance of a positive Developer Experience (DX) in enhancing engineering efficiency and confidence, noting that scaling DX can be complex due to disparate pipelines and tools. Leadership discussions highlighted the gradual journey toward instilling confidence in teams, especially within the DevOps space, where cultural shifts take time. The conference also addressed the challenges of cloud cost optimization through FinOps, advocating for shared ownership and cross-disciplinary collaboration akin to DevOps practices. Experimentation in software delivery, facilitated by Progressive Delivery and Feature Flags, was another key focus, with Harness announcing tools to streamline feature management and reduce failure rates. Furthermore, the Harness Platform introduced Test Intelligence to optimize test coverage and execution times, and AutoStopping/AutoStarting features to efficiently manage cloud resources without manual intervention. Overall, the event underscored Harness's commitment to democratizing software delivery by alleviating bottlenecks and enhancing agility in the software development lifecycle.

Harness Cloud Cost Management has introduced Cloud AutoStopping, a feature designed to significantly optimize cloud cost management by actively managing idle cloud resources, potentially saving users up to 75% on non-production environments. Traditionally, managing cloud costs has been challenging, often requiring manual intervention to implement cost-saving recommendations, which can lead to inefficiencies and unutilized savings. Cloud AutoStopping automates this process by detecting idle resources and either shutting them down or terminating them, while also allowing for automatic restarts when needed. This dynamic solution offers more efficient use of resources compared to static resource schedulers, eliminating access issues and reducing barriers to adoption by allowing teams to connect their infrastructure once without needing ongoing manual monitoring. Users can achieve substantial savings without the need for extensive learning or continual intervention, as highlighted by positive feedback from customers like Dheemanth R, CTO of Discover Dollar, who reported significant reductions in cloud expenses. Cloud AutoStopping is now accessible to Harness Cloud Cost Management users, offering an intuitive setup process to help organizations maximize their cloud cost efficiency.

Harness CI Enterprise (CIE) transforms the Continuous Integration/Continuous Deployment (CI/CD) workflow by introducing features like Test Intelligence, an integrated platform, and containerized steps, aiming to enhance developer efficiency and reduce build and testing times. Traditional CI tools, such as Jenkins, often involve manual scripting and lack robust testing, integration, and governance features. In contrast, Harness CIE, drawing from the open-source tool Drone, offers a modern solution with machine learning-driven Test Intelligence to optimize testing, a unified platform for seamless workflow management, and containerization to eliminate dependency issues. The platform supports GitOps, integrates with popular source control systems, and emphasizes security through features like encryption and fine-grained role-based access control (RBAC). Additionally, it offers a visual pipeline builder for easier sequencing and a developer-friendly interface, focusing on improving the overall developer experience by minimizing time spent on building, testing, and deploying code.

The {unscripted} Conference emphasized the importance of confidence in software delivery, focusing on modern Continuous Integration practices, iteration, and observability. Day one featured industry leaders like Dave Farley, who underscored confidence as a core theme for CI/CD pipelines, and discussions on the evolving nature of Continuous Integration and test coverage practices. Iteration was highlighted as key to building confidence in software engineering, with observability playing a crucial role in measuring and improving systems. Despite sophisticated observability practices, incidents can still occur, but effective incident management can mitigate their impact, as discussed by Robert Ross, CEO of FireHydrant. The discussions set a promising stage for day two, with further learnings and announcements anticipated from Harness.

Split has introduced a new Share Results feature that allows users to download and share experiment metrics in various formats such as PDF, CSV, and JSON, fostering improved communication and facilitating informed decision-making within teams. This tool aims to enhance experimentation culture by enabling team members to share insights, maintain comprehensive records, and conduct further analyses. The PDF format is suited for executive stakeholders with its visually appealing layout, while CSV and JSON cater to in-depth analysis and integration into internal systems. The feature is a response to customer feedback requesting better ways to convey experiment outcomes and build internal knowledge banks. Split also offers a range of learning resources through Split Arcade, which provides tutorials and challenges to help teams effectively manage feature releases and experiments. Users are encouraged to share feedback and connect with the Split team for support and improvements.

Harness Software Engineering Insights offers a streamlined solution for identifying and prioritizing code hotspots by correlating data from various DevOps tools, enhancing team efficiency in addressing areas that impede productivity and software quality. Code hotspots, akin to traffic slowdowns on highways, are sections of code where issues may not be immediately apparent but lead to frequent defects and increased developer time, thus affecting overall software performance and user experience. While redesigning or rewriting problematic code is ideal, it can be challenging due to dependencies or legacy code, making test automation a crucial strategy to manage these hotspots. Identifying these areas requires robust development metrics and data-driven decisions, which can be difficult and time-consuming with conventional methods like Excel or manual data correlation from tools like Zendesk, Jira, and GitHub. Harness Software Engineering Insights simplifies this process by automatically correlating data across platforms, providing clear insights into the impact and severity of code hotspots, and aiding in the maintenance of software quality and development speed.

Drone's evolution from a personal project to a significant player in the continuous integration (CI) space underscores the shifting landscape towards more accessible and enterprise-ready CI tools, challenging the dominance of Jenkins. Founded by Brad Rydzewski, who initially faced barriers with Jenkins during his tenure at General Electric, Drone began as a self-service CI platform aimed at simplifying software delivery. After transitioning from a SaaS product to an open-source model, Drone gained substantial popularity and contributions from major tech companies, leading to growing demands and Brad's eventual burnout. This prompted the introduction of paid support tiers and ultimately a strategic partnership with Harness, which shared Drone's vision of achieving Jenkins-level ubiquity in the CI marketplace.