April 2021 Summaries

Harness has introduced several enhancements to its Continuous Delivery platform, including delegate selection, pipeline execution details, and Azure GovCloud support, while also releasing a CI/CD Buyer's Guide to aid companies in the research and acquisition phases of CI/CD solutions. The platform now offers APIs for tag management and a Recommendations API for programmatically gathering insights from Kubernetes clusters. Harness has also been active in the media, with its cloud momentum announcement receiving coverage from various tech outlets, and it continues to engage with the community through podcasts like ShipTalk and participation in upcoming industry events such as KubeCon Europe and AWS Summits. Looking ahead, Harness is preparing for significant announcements at the upcoming {unscripted} conference, hinting at a major multi-module release that promises to further distinguish its offerings in the competitive CI/CD landscape.

Harness's Cloud Cost Management (CCM) platform provides businesses with tools to optimize Kubernetes workloads and clusters, aiming to achieve significant cost savings and improve operational efficiency. By utilizing the Recommendations API, organizations can automate the analysis and implementation of cost-saving strategies, ensuring optimal resource utilization. Users can leverage GraphQL queries to gather insights across Kubernetes clusters, fine-tuning recommendations based on specific sampling percentiles to achieve either cost or performance optimization. The platform assists in not just rightsizing workloads, but also offers guidance on resizing Kubernetes clusters for better efficiency. Harness CCM supports ongoing optimization efforts by providing a detailed analysis of resource usage and trends over time, helping organizations to make informed decisions about resource allocation and infrastructure scaling. Through the API Explorer, users can execute queries and apply recommendations to adjust and optimize their workloads and clusters, positioning Harness as a key partner in managing cloud costs effectively.

Reducing cycle time, from code commit to deployment, is crucial for enhancing engineering velocity and team satisfaction, with feature flags playing a significant role in this process by allowing code delivery and feature release to be decoupled. This separation enables teams to implement safer and more efficient iterations, resulting in smaller pull requests and faster reviews. Companies that focus on reducing cycle time typically report higher employee satisfaction and retention. The time between code commitment and production release, or cycle time, can be improved by measuring it as a Service Level Objective (SLO) and utilizing strategies such as limiting multitasking and reducing pull request sizes. Feature flags contribute to reducing cycle time by improving collaboration and allowing for incremental changes, which facilitate quicker reviews and approvals. Investing in continuous integration tools and automation can further reduce the time to deploy by increasing confidence in the deployment process. The article underscores the importance of balancing speed and thoroughness in reviews and suggests that feature management solutions like Split can enhance feature release speed and safety by providing better visibility and data-driven insights.

The Software Development Life Cycle (SDLC) is a framework comprising phases such as requirements gathering, design, development, testing, deployment, and maintenance, all crucial for delivering high-quality software efficiently. Organizations spend significant resources on deployment efforts, with models like Agile and DevOps reducing delivery times and enhancing team collaboration. The SDLC stages involve detailed planning, design decisions, building, testing, and operationalization to ensure software quality and performance. Various models have emerged over time, including the waterfall, iterative, and Agile models, each addressing different challenges and offering unique approaches to software delivery. Understanding the SDLC not only aids in improving delivery speed and quality but also enhances DevOps performance, incorporating feedback and insights to continuously deliver value. As the software industry evolves, the SDLC adapts to changing tools, processes, and demands, empowering teams to manage development complexity effectively.

Ancestry, a leading company in family history and consumer genomics, significantly improved its software deployment process by adopting Harness, achieving an 85% reduction in effort compared to its previous use of Jenkins. This transition enabled Ancestry to onboard 350 systems in a year, increase deployment velocity by threefold, and reduce production outages by 50%, with a 97% adoption rate of centralized testing tools. Initially, Ancestry's deployment process was inconsistent and unreliable, with different teams using custom Jenkins pipelines that led to downtime and operational challenges. Chief Architect Russ Barnett spearheaded the standardization efforts, focusing on implementing canary deployments and centralized testing to address these issues. Despite initial skepticism, the switch to Harness has been well-received due to its user-friendly interface and improved governance, leading to greater stability and productivity across Ancestry's platform.

James Bohrman's transition from an emergency dispatch technician to a DevOps engineer and Site Reliability Engineer (SRE) exemplifies the impact of storytelling and personal branding in technology careers. His journey emphasizes the importance of communication, continuous learning, and leveraging neurodiversity as a strength. Bohrman, who founded Cloudspeakers, utilized self-education and blogging to facilitate his career shift, showcasing how personal projects can significantly enhance professional visibility and opportunities. He advocates for embracing one's unique traits, especially for those who are neurodiverse, and underscores the value of listening and empathy in both personal development and technological environments. His insights also highlight the potential of content creation, such as blogging and video production, as tools for learning and community engagement in the tech industry.

CircleCI, a platform founded in 2011, is known for enhancing engineering team productivity and product quality by providing automation for software builds, tests, and deployments, handling over a million application builds daily. It uses a RESTful API and a config.yml file to define CI/CD processes, supporting modern workloads with features like Docker support, SSH debugging, and robust version control integration. The blog also evaluates alternative CI/CD platforms such as Jenkins, GitLab, and Harness. Jenkins is highlighted for its extensive plugin library and self-hosting capabilities, making it a flexible and reliable option, though lacking in version control for configurations. GitLab offers integrated end-to-end SDLC capabilities, including security dashboarding and built-in Kubernetes deployment, which is advantageous for users already in the GitLab ecosystem. Harness stands out with its comprehensive modules for CI, CD, and Cloud Cost Management, as well as numerous out-of-the-box integrations with various SDLC tools, offering both SaaS and on-prem solutions. Each platform presents unique benefits, and the post encourages a thorough evaluation based on specific organizational needs and challenges in software delivery.

Continuous Integration (CI) Testing is a crucial component of software development, ensuring quality and consistency by running automated tests during the build, packaging, and publishing phases. This process includes various tests, such as code quality, unit, integration, and security tests, to catch issues early and enhance confidence in software deployments. CI Testing allows for repeatability and rapid feedback, serving as quality gates that prevent progression if minimum quality standards are not met. It also distinguishes between on process and off process testing, where the former involves tests executed directly by the CI process and the latter involves external evaluations, such as container scanning. The use of CI tools like CircleCI and Jenkins facilitates this testing, and adopting a robust CI platform like Harness can further streamline the process by resolving dependencies and providing visibility into test coverage. As development practices evolve, the importance of an adaptable CI Testing framework becomes increasingly significant to support agile and microservices development, ensuring that software artifacts meet high-quality standards before deployment.

AppDynamics and New Relic are prominent Application Performance Management (APM) tools with distinct focuses, catering to different market segments. AppDynamics excels in providing end-to-end stack visualization and dynamic baselines, making it suitable for enterprises, while New Relic prioritizes response time metrics, appealing to startups and small to medium-sized businesses. Both originated from Wily Technology and share features like backend, mobile, and frontend monitoring, but differ in their approach to alerting and reporting, with New Relic using the Apdex score and AppDynamics relying on dynamic baselines. AppDynamics offers deeper database insights and on-premise solutions, whereas New Relic provides a more startup-friendly pricing model and supports additional programming languages like Ruby and Python. Both tools enable integration with various services and offer plugins, but New Relic's integrations are considered richer. Ultimately, the choice between them depends on the specific needs of the business, with AppDynamics being more suited for enterprises seeking customized solutions and New Relic being more accessible for startups with its transparent pricing.

Implementing DORA's four key metrics—Lead Time, Deployment Frequency, Mean Time to Restore (MTTR), and Change Failure Rate—enables organizations to effectively measure and enhance their DevOps performance, leading to faster and more stable software delivery. High-performing teams achieve a Change Failure Rate between 0-15%, showcasing the effectiveness of these metrics in fostering continuous improvement. The DORA team, a Google research group, has been pivotal in understanding DevOps practices over a six-year study, with their findings published in the State of DevOps Report and the book "Accelerate," co-authored by Nicole Forsgren. The book emphasizes the importance of tracking team outcomes rather than individual outputs to build high-performing technology organizations. These metrics provide insights into the tempo and quality of software delivery and are vital for organizations to discuss improvements and ensure shared responsibility for software quality. Tools like Harness' Continuous Insights and the open-source project Four Keys offer platforms to track and analyze these metrics, aiding organizations in their journey toward enhanced DevOps capabilities.

Kustomize and Helm are pivotal tools for managing Kubernetes configurations and packages, each offering unique benefits to streamline deployment processes. While Kustomize focuses on configuration management by allowing the refactoring of Kubernetes YAML manifests without additional dependencies, Helm serves as the de facto package manager for Kubernetes, facilitating the installation and uninstallation of software through Helm Charts. Helm provides a more comprehensive ecosystem with repository support, yet it has faced criticisms, particularly with its earlier versions due to security concerns related to Tiller. Additionally, alternatives like Jsonnet and Skaffold provide other templating solutions, each with distinct features and approaches. Harness enhances the efficiency of Kubernetes deployments by supporting both Kustomize and Helm, simplifying the deployment process for users regardless of their expertise level.

Split's platform enhances experimentation velocity by enabling simultaneous A/B tests, allowing companies to optimize features and accelerate product delivery. The key to effective parallel experimentation lies in the randomization of user groups to ensure independent impact distribution, thus maintaining accurate results. Running concurrent experiments maximizes development speed and flexibility, as it allows for testing and feature rollouts without waiting for sequential experiment completions. However, attention must be paid to potential interaction effects when two changes might interfere with each other, necessitating careful review and testing. Split emphasizes the importance of feature flags and experimentation as essential tools for software development, offering resources and support to help organizations leverage these capabilities for business impact.