March 2021 Summaries

Raisin, a leading pan-European wealth management platform, successfully transitioned from using Jenkins to Harness for software deployments, resulting in a 60% reduction in deployment effort and $525,000 in annual savings. The change increased deployment frequency by fivefold, enabling 53 deployments per month and significantly reducing new application onboarding time from two days to just 30 minutes. Initially, Raisin's central DevOps team, led by José Meyer, faced challenges with Jenkins due to its complexity and high maintenance demands, which consumed 80% of the team's time. The switch to Harness allowed the distribution of software deployment responsibilities to application teams, democratizing and automating the process, and reducing the DevOps team's involvement to 20% of their time. The streamlined process and improved efficiency with Harness facilitated quicker application deployment and attracted interest from various application teams within the company.

Error tracking tools, such as OverOps, Rollbar, Raygun, Sentry, Airbrake, Bugsnag, and StackHunter, are crucial for identifying and resolving application errors to enhance reliability and user experience. These tools offer various features, such as actionable insights, error frequency tracking, and integrations with platforms like Slack, JIRA, and Datadog, to streamline error management processes. OverOps provides code-level insights for Java environments, while Rollbar and Raygun offer support for multiple languages and frameworks. Sentry and Airbrake focus on capturing errors in real-time and providing detailed dashboards, whereas Bugsnag and StackHunter present specialized solutions for error monitoring and analysis. The choice of tool depends on specific application needs, supported languages, and the workflow improvements desired, with many offering free trials for evaluation.

The text discusses various alternatives to Argo CD, an open-source Continuous Deployment solution for Kubernetes, highlighting tools like Flux, Jenkins X, GitLab CI/CD, and Harness CD, each offering distinct features and advantages for modern DevOps practices. Argo CD uses GitOps workflows to automate Kubernetes service deployments, ensuring the desired state of resources matches the current state. Flux offers GitOps capabilities with its GitOps Toolkit, emphasizing a declarative approach to system configuration. Jenkins X supports CI/CD for Kubernetes with extensive plugin options and GitOps integration, while GitLab CI/CD provides an all-in-one DevOps platform with robust GitOps support and a wide range of extensions for deployment strategies. Harness CD stands out for its ease of use, AI/ML capabilities, and comprehensive software delivery platform that includes CI/CD, cloud cost management, and feature flags, supporting both cloud-native and traditional workloads. Each tool offers unique customization options, security features, and support for Kubernetes-native environments, catering to diverse organizational needs in continuous delivery.

Harness has announced major updates across its products, including enhancements to its CI Community and Essentials editions with new UI improvements, pipeline execution visualization, and real-time debugging tools. Additionally, Harness introduced new features for its Cloud Cost Management and CD products, emphasizing its commitment to innovation and community engagement. The company also expanded its ShipTalk podcast series, which explores various aspects of DevOps with industry experts. Notable updates include the release of Drone 2.0, offering significant improvements for open-source CI solutions, and expanded support for Azure deployments in Harness CD. Harness Cloud Cost Management now includes visibility for Microsoft Azure, complementing existing support for AWS and Google Cloud. The industry-first "What If" analysis tool for Kubernetes cost savings recommendations is also highlighted, providing users with data validation and optimization options. These developments have generated significant media attention, with coverage from major outlets like TechCrunch and VentureBeat. Looking ahead, Harness hints at further cloud support announcements and enhancements to its Cloud Cost Management offerings.

Transitioning from Jenkins to Harness offers significant improvements in continuous integration (CI) by reducing the need for extensive scripting and manual configurations, leading to faster and more reliable software delivery. Jenkins, although a pioneering CI tool, is criticized for its outdated approach that requires manual intervention and complex scripting, which can be cumbersome and error-prone. The author shares personal experiences with Jenkins, highlighting the challenges of setting up pipelines and troubleshooting errors due to its dependence on plugins and manual configurations. Harness, on the other hand, is depicted as a modern alternative that streamlines the onboarding process, offering intuitive interfaces and automation capabilities that align better with contemporary cloud infrastructures. The platform's ability to simplify the creation of applications, services, environments, and workflows without the need for detailed Jenkinsfiles underscores its advantage in the evolving landscape of software development tools.

Harness CD offers a robust Continuous Delivery solution that addresses the limitations of Spinnaker, providing features like auto-generated deployment scripts, machine learning-based Continuous Verification, and deep integrations. Spinnaker, originally developed by Netflix and released as open source, is praised for its multi-cloud deployment capabilities but criticized for its lack of support for traditional applications and secrets management. The text evaluates various Spinnaker alternatives, including Harness CD, Argo CD, GitLab, Azure DevOps, Octopus Deploy, and Armory, each with unique features and drawbacks. For example, Argo CD excels in GitOps functionality but has a steep learning curve, while GitLab offers a comprehensive DevOps platform but lacks native secrets management. Azure DevOps provides flexible pipelines but requires significant scripting, whereas Octopus Deploy is easy to use but lacks advanced features. Armory extends Spinnaker's capabilities with additional modules but at a high cost. The text suggests considering these alternatives based on specific needs and encourages trying Harness, emphasizing its scalability and comprehensive features.

Harness has significantly enhanced the Drone project by introducing a new user experience, a CI pipeline visualizer, a real-time debugger, and user management features, effectively doubling its engineering investment and providing enterprise support to improve continuous integration for developers. These advancements, which come just over six months after Drone joined the Harness family, are aimed at simplifying the developer self-service experience for continuous integration. The upgrades include a revamped interface to create a seamless experience for users of both Drone for CI and Harness for CD, an improved dashboard with build times and charts for better visibility of build failures, and a graph view for visualizing complex CI pipelines. Additionally, a real-time debugger allows developers to quickly inspect and debug failed pipelines, while new user management screens simplify user oversight, with upcoming enhancements planned for role-based access control and governance.

In a ShipTalk podcast episode, DevOps Engineer Donovon Carter from Dexcom discusses the challenges and insights of balancing innovation with regulatory requirements in the highly regulated medical device industry. Carter's non-traditional journey from a political science major to a DevOps role highlights the importance of interpersonal skills and negotiation in technology fields. At Dexcom, a leader in continuous glucose monitoring, Carter emphasizes the critical role of automation and continuous integration/deployment practices in maintaining compliance while fostering development speed. Despite the constraints of stringent regulations, innovation thrives through careful iteration and feedback within controlled environments, ensuring safety and quality before reaching customers. Carter also reflects on industry trends, advocating for simplicity in design and the reuse of well-established tools to enhance efficiency and reduce complexity in software development.

Tracking key DevOps metrics such as deployment frequency, lead time for changes, time to recover, and change failure rate is crucial for enhancing Continuous Integration (CI) processes, improving code quality, and accelerating delivery. The DevOps Research and Assessments (DORA) team has identified these four metrics as essential for understanding DevOps performance challenges and aligning with characteristics like speed, visibility, and reliability. Continuous Integration involves automating the building, testing, and validation of code changes, helping developers accelerate release cycles and reduce merge conflicts. Key performance metrics for improving CI include code quality, code coverage, test coverage, defect escape rate, and lead time to deploy. These metrics help teams identify areas for process improvement, encourage better testing practices, and streamline deployment, ultimately enhancing the overall CI/CD pipeline. Tools like Harness can provide actionable insights and automation to further refine these processes, making it easier to achieve high-quality, efficient software delivery.

Split is a feature management platform that combines feature flags with meaningful user events to enable effective experimentation and performance analysis in production environments. By integrating with tools like mParticle, Segment, or Google Analytics, Split customers can leverage existing events to enhance their data-driven decision-making processes. The platform supports event extraction through webhooks or REST APIs and provides guidance on transforming and loading events for better analytics. Split's integration with various data platforms allows for seamless event tracking, and it emphasizes the importance of measuring feature impact to avoid the "impact gap," where new features may have a neutral or negative effect. By offering detailed insights into feature performance, Split helps teams focus on innovations that truly resonate with customers, ensuring safer and more impactful feature releases.

Continuous delivery tools are essential for automating and standardizing the software deployment process, catering to diverse needs from small teams to large enterprises. The article explores a range of tools, from open-source options like Argo CD, which excels in automating Kubernetes deployments using GitOps, to comprehensive platforms like Harness, known for its governance and modular capabilities in CI/CD. It highlights the evolution of continuous delivery from its origins in the early 2000s, emphasizing the decoupling of CI from CD due to the cumbersome scripting required in earlier CI tools like Jenkins. The text categorizes CD tools into open-source, platform solutions, cloud provider tools, and CI tools with CD capabilities, noting each category's strengths and limitations. Moreover, it underscores the importance of integrating existing tools like code repositories and infrastructure automation tools into CD pipelines to enhance sophistication and efficiency. Ultimately, selecting the right tool depends on factors such as organizational needs, expertise, and the resources available for setup and maintenance.

Jim Shilts, founder of the North American DevOps Group (NADOG) and developer advocate at Shipa, shares his journey through the evolution of DevOps from its early stages to the current cloud-native era, emphasizing the significance of community involvement and the transition to Kubernetes and microservices. NADOG, initially started as a local community for IT practitioners to share knowledge and foster connections, expanded rapidly, adapting to virtual formats during the COVID-19 pandemic while maintaining its core values of learning and collaboration. Shilts reflects on the challenges and shifts in engineering efficiency over the years, describing how DevOps has moved from a focus on automation to a broader understanding that includes culture, collaboration, and breaking down silos between development and operations. He highlights the importance of community engagement and continuous learning, encouraging those new to the field to get involved in community groups like NADOG to benefit from shared experiences and networking opportunities.

Swallowed exceptions in Java applications pose significant challenges for developers, often leading to unnoticed errors that can negatively impact user experience and business metrics. These issues are exacerbated by the presence of empty catch blocks, where exceptions are not logged or handled, making it difficult to diagnose issues from log files. Research indicates that swallowed exceptions are as prevalent as logged ones, highlighting a need for improved error handling practices. To address this, developers are encouraged to refresh code review guidelines to eliminate empty catch blocks, refactor logging practices to include meaningful statements, and adopt continuous reliability measures to automate root cause analysis. These strategies aim to enhance application reliability by ensuring that errors are properly documented and analyzed, ultimately making swallowed exceptions a problem of the past.

SRE (Site Reliability Engineering) and DevOps, while sharing common goals of improving system reliability and operational efficiency, differ in their focus and approaches. SRE emphasizes maintaining system reliability through engineering practices, often involving development-focused engineers who tackle operational and scalability challenges. In contrast, DevOps centers on collaboration and automation between development and operations teams to enhance engineering efficiency, involving ops-focused engineers who address development pipeline issues. Despite their overlapping goals, they solve distinct problems, with SRE concentrating on system health and uptime, while DevOps targets removing bottlenecks in the software delivery lifecycle. Both roles use metrics, such as SLAs, SLOs, and SLIs, to assess system performance, but DevOps also measures software delivery performance through the Accelerate metrics. Harness, a platform for Continuous Delivery and Verification, supports both DevOps and SRE teams by providing tools to manage deployment, track software reliability metrics, and enhance system resilience through Chaos Engineering.

Patricia Anong's journey from psychology to becoming a self-taught DevOps consultant underscores the significance of adaptability, continuous learning, and networking in the tech industry. After feeling uncertain about her future in psychology, Patricia transitioned into tech, leveraging self-directed learning and curiosity to gain expertise in DevOps. Her career spans roles at companies like Under Armour and AT&T, and she shares her knowledge through a personal blog focused on CI/CD, best practices, and multi-cloud environments. Patricia emphasizes that success in tech requires a combination of formal education, self-learning, and hands-on experience, and she encourages women to pursue their goals despite the industry's male dominance. Her experiences reflect both the challenges and opportunities for women in tech, as she highlights the importance of passion, skill development, and mentorship in advancing one's career.