August 2023 Summaries

Containers and microservices have transformed cloud application deployment, with Kubernetes emerging as the standard for container orchestration since 2014. This tutorial guides users through deploying a Node.js application on Azure Kubernetes Service (AKS) using progressive delivery methods with CircleCI and Argo Rollouts, which facilitates controlled, gradual feature releases to reduce risk. It involves setting up a CI/CD pipeline that automates blue-green deployments, allowing developers to release updates to a small user subset initially, thereby identifying issues before a broader rollout. Argo Rollouts, built on Kubernetes API, supports advanced deployment strategies such as canary and blue-green deployments, minimizing downtime and risk by enabling seamless traffic switching between application versions. The tutorial also covers prerequisites, application containerization, Kubernetes manifest configuration, AKS cluster creation, and ArgoCD installation for deployment management. It emphasizes GitOps principles, using YAML for configuration, and explains manual and automatic sync policies for application deployment. The CI workflow integrates CircleCI orbs to streamline processes, and the tutorial concludes with guidance on monitoring applications via the ArgoCD dashboard, promoting new application versions, and performing rollbacks if necessary.

Feature flags, also known as feature toggles or switches, are integral to modern software development, allowing developers to control the visibility and accessibility of features within applications. These tools are vital for managing feature releases in continuous deployment (CD) environments, where agility and automation are crucial. Feature flags help mitigate risks and enhance user satisfaction by decoupling feature releases from code deployments, enabling rapid and controlled rollouts, A/B testing, and swift rollbacks when necessary. Various types of feature flags, such as release, operational, experimental, and access control flags, offer developers flexibility and control over feature management, ensuring stable and reliable software delivery. However, their implementation requires careful planning to avoid challenges like flag proliferation and testing complexity. By incorporating feature flags into CD practices, development teams can deliver bug-free, stable software more quickly and efficiently, leveraging platforms like CircleCI to enhance their deployment pipelines.

The CircleCI's annual State of Software Delivery report reveals key trends and insights in programming languages and their impact on software delivery performance, based on data from over 14 million workflows. As CI/CD and DevOps become integral to organizations aiming for efficient delivery cycles, programming language choices influence developer productivity and performance metrics such as workflow duration, throughput, mean time to recovery, and success rate. The report highlights the growing popularity of languages like Python, TypeScript, and C# due to their support for rapid, scalable, and secure development across various applications, including machine learning, enterprise systems, and cross-platform development. Python has surged in usage thanks to its versatility and powerful libraries for AI and data science, while TypeScript has maintained its dominance over JavaScript for building complex applications. C# continues to gain traction for its cross-platform capabilities and integration with Microsoft's ecosystem. This data underscores a shift towards languages that offer a blend of speed, security, and scalability, contributing to faster error recovery and improved software delivery performance.

Platform teams play a crucial role in the success of world-class software engineering organizations by enhancing developer productivity, creating streamlined pathways for team success, and automating solutions. These teams enable developers to concentrate on essential tasks and significantly contribute to the success of research and development efforts. CircleCI acknowledges the importance of platform teams and celebrates the contributions of various leaders in the field, such as those from Tinder, Iterable, Top Hat, and other prominent companies. These individuals are recognized for their impact on the platform space, demonstrating how this emerging discipline is integral to scaling engineering operations effectively.

Organizations aiming for effective long-term machine learning (ML) investments are creating custom workflows tailored to their specific data and outcomes, with an emphasis on keeping models updated and trained on the latest data. This tutorial illustrates how to automate ML workflows using CircleCI, breaking them down into steps for continuous integration (CI) and continuous deployment (CD), which include building, training, testing, deploying, and retraining models. It highlights how CircleCI's platform facilitates automation, allowing ML models to be updated and validated automatically, thereby freeing experts from manual tasks and enabling them to focus on developing more accurate models. Additionally, it discusses the integration of GPU resources for efficient model training and the importance of MLOps strategies to ensure the models remain effective through regular updates and testing. The tutorial also mentions the use of CircleCI's web console for monitoring workflows and approving model updates, and the flexibility it provides in configuring custom notifications and leveraging both cloud and on-site processing power.