July 2020 Summaries

In a detailed guide, a method is presented for creating a basic CRUD application using Spring Boot and Angular, underscoring the significance of feature flags for enhancing software deployment. The guide explains how to set up the development environment, create a backend using Spring Boot with a focus on REST endpoints, and integrate a front-end Angular application. It emphasizes the use of feature flags, enabled through the Split platform, to facilitate continuous deployment and safe feature releases, allowing developers to test in production and foster a culture of experimentation. The tutorial includes steps for incorporating feature flags to add a delete function to the app, demonstrating both backend and frontend integration, and highlights the benefits of feature flags in reducing release risk and supporting DevOps best practices. Additionally, resources for further learning about Spring Boot, Angular, and feature flag testing are suggested, along with an invitation to explore Split's feature management platform to enhance software development processes.

Harness has introduced several new features aimed at enhancing Cloud Cost Management and operational efficiency, including the ability to export cost and usage data via their GraphQL API and deployment log exports from both the UI and API. These updates are designed to help users optimize their cloud expenditure, with additional improvements such as Windows Remote Management authentication via Kerberos and the ability to destroy remote Terraform Backends. Harness is also hosting its first industry conference, {Unscripted}, on October 21st-22nd, inviting participation and paper submissions until August 17th. The company has been gaining attention for its achievements, such as 675% global partner growth, and offers virtual events on topics like microservices and agile foundations, while also promoting their demo sign-ups with a free T-shirt incentive.

Transitioning from an Integrated Development Environment (IDE) to command-line tools and custom scripts can enhance system performance, reliability, and flexibility, particularly for developers managing multiple environments and remote servers. The author shares their personal journey of deleting their IDE due to performance issues and inefficiencies, advocating for a command-line approach that offers deeper control over tasks like compiling, linking, and publishing code across various programming languages. This method is especially beneficial for technical leads and developers frequently switching tasks or working on operations tasks, as it minimizes the resource strain and interference associated with running multiple IDEs. The transition requires a solid understanding of shell environments and tools like Bash, Gradle, and CMake, but it empowers developers to automate complex tasks, improve debugging with debug prints, and optimize workflows with custom scripts. While the learning curve is steep and not recommended for beginners, developers frustrated by their IDE's limitations may find this approach rewarding.

Branch by abstraction is a software development technique that facilitates continuous deployment by allowing large code changes to be made gradually and safely through an abstraction layer. This pattern is crucial for trunk-based development and continuous integration and continuous delivery (CI/CD) pipelines, as it ensures that code alterations can coexist with ongoing development and that the entire codebase remains releasable at any time. The process involves creating an abstraction layer that enables communication between new and old systems, allowing for incremental updates without disrupting existing functionality. Feature flags can enhance this process by allowing developers to toggle access to new code, conduct limited releases, and safely test in production environments. This method reduces merge conflicts and allows for easy pausing and resuming of changes, although it may not be suitable for scenarios where customers choose their own upgrade timelines. Tools like Harness can further streamline this process by offering AI-driven automation and feature management capabilities, enabling faster and safer code deployments.

Automating the delivery pipeline using the Harness platform significantly enhances the speed and efficiency of software deployments, making the continuous integration and continuous delivery (CI/CD) process more seamless. Ravi Lachhman, a CI/CD Evangelist, provides insights and a quick demonstration on how to achieve this automation, showcasing the capabilities of the Harness platform in transforming CI/CD operations.

Kubernetes has revolutionized workload scaling by providing a platform where application and infrastructure teams can unify their deployment processes using simple YAML configurations, though it has introduced challenges in cost management due to its ease of scaling. Harness Cloud Cost Management offers tools to optimize and track Kubernetes cluster expenses, highlighting the correlation between deployment changes and cost impacts. Before Kubernetes, scaling distributed applications required complex processes involving middleware engineers and was hindered by infrastructure constraints. Now, Kubernetes allows for rapid scaling through commands and YAML adjustments, reducing the need for detailed infrastructure knowledge but increasing potential costs as non-production workloads can consume significant resources. Platform engineers focus on maintaining the Kubernetes environment, while application teams handle workload tuning. To address cost concerns, Cloud Cost Management by Harness provides insights into application and cluster usage, enabling better financial oversight and optimization strategies without relying on traditional labeling or tagging methods.

The guide details a step-by-step approach to building a basic Spring Boot web application that displays weather information for Detroit, Michigan, while incorporating feature flags using Split.io. It begins by setting up necessary accounts and dependencies, including Split's Java SDK and the Open Weather Map API, before using the Spring Initializr to create the application's scaffolding. The process involves creating a simple HTML page to display weather data and implementing a controller to manage data retrieval, initially using hardcoded values, and then toggling between a stub service and a live weather service via feature flags. The use of feature flags offers dynamic switching between different implementations and introduces basic rules for split management, demonstrating how to enable or disable features seamlessly. The guide encourages further enhancement of the application by suggesting additional features like input fields for different locations, unit tests, advanced split treatments, and styling improvements, while also highlighting the broader potential of feature flags in application development.

Harness significantly reduced its Google Cloud Platform (GCP) logging costs by identifying and minimizing unnecessary log volumes, implementing log exclusions, and redirecting logs to more cost-effective storage solutions, leading to a 50% reduction in daily logging expenses and annual savings of over $140,000. Initially, the logging costs were around $800 per day, largely driven by log volumes from Google Kubernetes Engine (GKE) and Cloud HTTP Load Balancer. To better understand and manage these costs, Harness utilized tools like BigQuery for detailed log analysis, structured logs into JSON payloads for enhanced querying, and attributed costs to specific code areas contributing to log noise. By configuring log exclusions and reducing log volumes from particularly noisy components, the company was able to optimize its cloud logging expenses effectively. This approach not only achieved significant cost savings but also aligns with Harness's commitment to continuous cloud cost management.