November 2022 Summaries

Dimensional analysis in Split's platform enhances experiment optimization by organizing data into valuable dimensions, allowing businesses to assess the impact of new features on performance metrics. This approach helps companies understand unexpected outcomes and improve their experimental processes by utilizing event property data across multiple sources. By configuring dimensions, Split periodically reviews event data streams and calculates metrics based on unique property values, offering insights into user behavior and conversion rates. Practical examples, such as one-click purchases and Apple Pay integration, demonstrate how dimensional analysis can be applied to A/B testing, emphasizing the importance of selecting relevant dimensions like platform, browser, and engagement level. The guide stresses the need for collaboration among analytics, data engineering, and user experience teams to ensure effective experimentation and encourages businesses to think critically about dimension selection to derive meaningful insights. Split's Feature Data Platform allows for safe feature deployment through feature flagging, enabling fast, reliable experimentation and decision-making.

Harness Cluster Orchestrator for Amazon EKS, a feature within the Harness Cloud Cost Management (CCM) module, empowers engineering, DevOps, and CloudOps teams to efficiently manage EKS clusters by providing workload-driven intelligent autoscaling, cost visibility, and automated optimization. By leveraging AWS Spot Instances and the distributed Spot orchestration capability, users can achieve up to 90% savings on cloud costs. The tool offers a simplified approach to workload and node resizing, ensuring that clusters are not over-provisioned, and provides options to balance Spot and On-Demand nodes for critical workloads. Harness, recognized as part of the AWS Service Ready Program for Amazon EC2 Spot instances, supports best practices and API integration for successful management of Spot instances. The Cluster Orchestrator also offers comprehensive features including workload distribution strategies, cost visibility down to an hourly level, anomaly detection, and workload resizing recommendations to optimize resource use and reduce expenses.

Harness Feature Flags has been recognized as a leader and outperformer in the GigaOm Radar Report for Feature Flags, highlighting its advanced capabilities in feature management, including no-code workflow customization, policy-based governance via Open Policy Agent (OPA), and seamless CI/CD integration. Launched just a year ago, this platform aims to enhance the speed and safety of feature releases by offering a balanced solution that supports both rapid development and robust governance. Harness Feature Flags is noted for its innovative approach, enabling Dev and DevOps teams to automate feature releases through user-friendly, no-code pipelines and a config-as-code model with GitOps, thus facilitating better collaboration and reducing manual management. By integrating feature management with CI/CD, Harness provides end-to-end software delivery support, allowing teams to focus on more customer-centric rollouts and make informed business decisions. The platform's ability to streamline the release process, minimize errors, and improve audit capabilities is setting it apart in the rapidly evolving feature flag market.

Migrating from Jenkins to Drone CI offers a more modern, efficient, and container-based CI/CD solution, ideal for organizations adopting cloud-native technologies and GitOps methodologies. Drone CI, built on Docker and utilizing declarative YAML configurations, simplifies setup and maintenance, making it suitable for both developers and operations engineers. Unlike Jenkins, Drone CI is designed for the modern cloud-native world, providing advantages such as faster build pipelines, easier scalability, and reduced maintenance time. Jenkins, while historically popular due to its extensive plugin ecosystem, struggles with scalability, high maintenance, and lacks the modern capabilities required by enterprises today. Drone CI, written in Go and leveraging Docker containers, facilitates a more streamlined and agile CI/CD process and is particularly advantageous for organizations prioritizing simplicity, speed, and a modern approach to software deployment.

Harness CI, founded by Brad Rydzewski, demonstrates significant performance enhancements with build times up to four times faster than competitors like GitHub Actions and other CI tools. Key features such as Cache Intelligence and the exclusive Test Intelligence™ contribute to these improvements, with Test Intelligence using machine learning to optimize which tests are necessary, significantly reducing the number of tests and running them in parallel. This results in substantial time and cost savings, offering up to 300% annual savings on infrastructure costs. The performance was tested using popular open-source projects like Apache Kafka, Apache RocketMQ, and Apache Zookeeper, showing remarkable speed advantages and validating Harness CI's commitment to improving the developer experience by removing friction and enhancing efficiency.

Microservices architecture facilitates scalable and independent service deployment, enhancing development agility and system resilience by allowing developers to create smaller, standalone software services that can be updated more frequently. These services, which are lightweight, stateless, and modular, enable teams to update individual components on their own schedule, improving responsiveness to user needs. The tutorial demonstrates deploying a Node.js microservice on Kubernetes using Harness, a modern software delivery platform, which simplifies the deployment process by leveraging Docker and Kubernetes for scalability and continuous integration. Microservices are ideal for large, complex applications that require scalability and are often used alongside containers and service-oriented architecture, but their benefits can vary depending on the application's size and implementation strategy.

Multi-stage Docker builds are a technique to optimize Docker images by breaking the build process into multiple stages, significantly reducing image size and improving security by minimizing dependencies. This method streamlines the deployment process by creating images that only include necessary components, thereby saving time and reducing the risk of vulnerabilities. For example, a Node.js application saw a reduction from 48.81 MB to 7.12 MB when utilizing this approach. Multi-stage builds allow for the separation of build and runtime environments within a single Dockerfile, facilitating the management of different environments such as development, staging, and production without the need for multiple Dockerfiles. This technique not only enhances efficiency by speeding up build times but also simplifies debugging and allows the reuse of cached images. Implementing multi-stage builds can lead to more efficient and secure software deployment.

Harness celebrates Veterans Day by spotlighting two veterans within their team, showcasing how their military experiences have equipped them for careers in technology and fostering a supportive community for veterans transitioning to civilian roles. The article features Michael Wolfe, a former Marine Corps Cyber Network Operator who joined Harness as a Senior Large Enterprise Sales Engineer, and David McNamara, a former U.S. Coast Guard member now serving as a Staff Database Reliability Engineer. Both veterans share insights into their military service, memorable experiences, and the skills they acquired that translate well into their current roles. Harness also emphasizes the importance of recognizing veterans' contributions and offers resources and benefits for veterans interested in the tech industry, including discounts and opportunities to join their team. The company expresses gratitude for the unique perspectives veterans bring and encourages them to explore career opportunities within Harness.

Harness Service Reliability Management (SRM) is a tool designed to enhance software reliability by capturing and analyzing code-level data throughout the software delivery lifecycle. It emphasizes Continuous Reliability (CR), which balances speed, complexity, and quality by proactively managing reliability. SRM captures various types of data, including code metrics, true root causes, transaction performance, and system metrics, and applies advanced machine learning for prioritization and anomaly detection. This comprehensive data collection helps organizations pinpoint critical gaps and prioritize issues, offering insights beyond traditional stack traces by providing detailed information such as the source code and environment variables at the moment of an incident. By correlating code-level failures with system metrics, SRM offers a holistic view of application performance and dependencies, which aids in identifying the root causes of issues. The tool's unique capabilities are attributed to its use of code mapping, runtime code analysis, software data optimization, and machine learning, ensuring efficient and detailed observability and reliability in software delivery.

Harness has integrated Apex into its Feature Flags platform, allowing Salesforce developers to utilize advanced feature management capabilities through a new open-source Apex SDK. This addition aims to streamline the integration process, enabling development teams to deliver updates more swiftly and securely. Apex, a popular language for Salesforce platform software, can now take advantage of feature flags to enhance user experiences and expedite the delivery of business-critical applications. The Apex SDK provides Salesforce apps the same robust feature management, fostering faster and safer software delivery and innovation. Developers can access detailed setup and deployment steps through the Apex SDK Reference, ensuring a straightforward implementation process.

Adopting a unified software development lifecycle (SDLC) platform like Harness can significantly enhance productivity and governance by consolidating processes and metrics into a single repository, thereby reducing tool sprawl, maintenance burdens, and security risks. Many Chief Technology Officers (CTOs) traditionally opt for a collection of point solutions to manage software delivery and DevOps, which can lead to increased complexity, integration costs, and fragmentation, as outlined by Metcalf's law. This fragmented approach not only demands a substantial portion of developers' time for maintenance but also creates dependency on individual knowledge, which can be risky if key personnel leave. By centralizing software delivery processes, an SDLC platform can mitigate these issues, offering a more cohesive, secure, and efficient development environment, and allowing teams to focus on innovation and customer satisfaction rather than internal complexities.