June 2023 Summaries

Harness has launched the Continuous Delivery & GitOps Administrator Certification, enhancing its certification framework by adding a practical hands-on component to validate skills and manage Continuous Delivery (CD) capabilities at scale. This new certification, which follows a developer-level certification, includes a combination of a knowledge-based exam and practical exercises using the Harness Continuous Delivery & GitOps Enterprise Module, accessible via a cloud shell without the need for personal infrastructure. The certification process is facilitated through the Harness Developer Hub, where candidates can register for the exam and access learning materials, with successful participants receiving a Credly Badge. Harness plans to introduce further certifications across different levels and modules to support continuous learning and skill enhancement.

Harness CI's Remote Debugger is an advanced feature that allows developers to engage directly with failed pipeline executions using SSH, facilitating real-time debugging and significantly reducing troubleshooting time. This tool enables developers to rerun pipelines in debug mode, providing a deep dive into the core issues by inspecting system and application logs, runtime environments, and network configurations directly within the CI stage where the problem occurred. It offers practical solutions for real-world scenarios such as debugging failed tests, investigating integration issues, and identifying performance bottlenecks. The Remote Debugger transforms the debugging process from a time-consuming and speculative task into a focused and efficient real-time investigation, fostering a deeper understanding of application behavior and contributing to the development of robust and high-quality software. Currently available as a preview feature, it represents a paradigm shift in continuous integration debugging by enhancing developer efficiency and supporting continuous improvement.

Split's new integration with Datadog RUM (Real User Monitoring) enhances product development by providing teams with crucial insights into user experience, allowing for more frequent and innovative releases without compromising quality. This collaboration overlays feature flag information on RUM data, enabling teams to monitor real-time user activity and quickly address issues by rolling back or shutting off problematic features. The integration allows for detailed feature flag analysis, including user exposure and error counts, and can break down information by dimensions like country and browser type. To utilize this integration, teams need the latest Datadog RUM Browser SDK and Split's SDK for feature flag evaluation reporting. Split positions itself as a comprehensive feature management and experimentation solution, facilitating faster, safer, and more insightful product releases while emphasizing continuous improvement and strong customer support.

Harness has introduced its AI Development Assistant (AIDA), a comprehensive AI-infused platform designed to revolutionize the software delivery lifecycle by automating tasks, enhancing security, and accelerating delivery processes. This initiative integrates AI across various stages, offering capabilities such as build troubleshooting, cloud asset policy generation, and security vulnerability remediation. Harness AI aims to boost productivity by automating CI/CD pipelines, code analysis, and code reviews, while also providing AI-assisted onboarding and test case generation. The platform addresses security concerns by automatically generating and enforcing policies, explaining vulnerabilities, and suggesting remediations. Despite the potential challenges in data privacy, system integration, and the build-vs-buy debate, Harness's history of leveraging AI/ML in DevOps positions it to overcome these barriers effectively. The company emphasizes a privacy-first approach, ensuring its AI tools operate within robust governance frameworks. This initiative marks a significant leap in DevOps, positioning Harness AI as a transformative force in modern software delivery.

Harness has been recognized in the GigaOm Radar for Cloud FinOps v2.0 for its innovative cloud cost management solutions, particularly its Cloud AutoStopping and Kubernetes support, which enable businesses to save significantly on non-production cloud infrastructure costs. In today's data-driven environments, managing cloud expenses is crucial as decentralized cloud usage can lead to inefficiencies and wasted resources. Industry analysts like GigaOm offer expert insights into available solutions, helping businesses make informed decisions. Harness, noted for its automation capabilities, is classified as a Challenger in the GigaOm report but is anticipated to advance to leader status due to its innovative approach to cost savings. The company's Cloud AutoStopping feature proactively manages idle resources, saving up to 80% on non-production infrastructure, while its Kubernetes support helps optimize cluster usage and reduce costs. Harness also enhances real-time decision-making with its Cost Perspectives and Cost Categories features, earning an Exceptional rating in this area from GigaOm.

Harness Continuous Verification utilizes machine learning to enhance deployment reliability by analyzing metrics and logs for potential regressions, enabling safer deployment strategies such as canary releases. This approach minimizes the risk of deployment failures by detecting and addressing anomalies without manual intervention. Continuous Verification validates deployments by comparing current and previous data, clustering related events, and calculating deviations to identify potential risks. The process involves metric and log analysis, where metrics provide time-series data and logs offer unstructured data, both analyzed for deviations and anomalies. This systematic safeguard ensures that deployments are robust and reliable, supporting continuous iteration and innovation in technology development while maintaining stability. Harness's methodology emphasizes the importance of understanding normalcy in deployment data and making informed judgment calls based on detailed analysis.

Setting up a threat model is a systematic approach involving the identification of security objectives, application decomposition, threat determination and ranking, and implementation of countermeasures to enhance application security. Security objectives guide threat modeling efforts by defining constraints and focusing on sensitive data protection. Application decomposition provides an understanding of interactions with external entities through structured information gathering, identifying roles, scenarios, technologies, security mechanisms, and data flow diagrams. Threats are categorized and analyzed using models like STRIDE and DREAD to assess risk factors, allowing for the prioritization of threats based on their likelihood and impact. Countermeasures are identified to mitigate vulnerabilities, and a threat modeling report is created to summarize findings and draw management's attention to security strategies. The process ultimately aims to reduce overall risk and can be integrated into workflows using modern DevSecOps solutions such as Harness STO.

Conversion rates are crucial for understanding how effectively a business meets its customers' needs, and feature flags offer a dynamic way to enhance these rates by enabling agile, data-driven optimizations. Traditionally used in software development to control feature visibility, feature flags can also serve as powerful marketing tools, allowing teams to test and refine user experiences to boost conversion performance. By facilitating collaboration between marketing and development teams, feature flags help tailor offerings to user preferences, enabling personalized experiences that resonate with different customer segments. This approach supports incremental changes, providing real-time insights into user behavior and engagement, and ensuring businesses remain adaptable and responsive to market trends. Platforms like Split empower companies to utilize feature flags effectively, offering solutions that allow for real-time configuration adjustments and fostering a partnership that supports ongoing improvement and innovation.

Split's localhost mode with JSON files enables developers to experiment with feature flags offline, enhancing productivity during internet interruptions or while traveling. This feature is accessible through Split's server-side SDKs for Java, Go, and Python, allowing users to download and locally store Split environment definitions and user segments in JSON format. By using the reserved word "localhost" in the Split API key, developers can direct the SDK to utilize local files, ensuring seamless feature flag evaluation even offline. The SDK can regularly poll for changes in these files, facilitating real-time updates and experimentation. This functionality caters to the growing digital nomad trend and provides flexibility to developers working in varying environments. Additionally, Split's Feature Data Platform supports feature management and experimentation, offering tools for A/B testing and controlled feature releases without sacrificing speed or reliability.

Harness has introduced its Internal Developer Portal (IDP) powered by Backstage, designed to streamline developer onboarding and service management by automating processes and reducing complexity. The IDP addresses common challenges faced by companies, such as lengthy onboarding times, fragmented technology choices, and the cognitive overhead of managing a decentralized, cloud-based microservices environment. With features like service onboarding pipelines, user management, and documentation integration, Harness aims to enhance developer productivity by allowing developers to focus on writing features while platform engineers manage standards and automation. The platform also includes a catalog of software components and a robust search functionality, enabling efficient access to technical documentation and service information. Additionally, the extensibility through a plugin architecture allows for integration with third-party providers, enhancing the portal's capabilities. Currently in a limited beta release, Harness is working closely with customers to refine the IDP for broader use cases.

Integrating chaos experiments into Harness CD pipelines can significantly enhance service reliability and developer efficiency by identifying design and architecture gaps early and reducing resilience debt through enforced resilience testing policies. This proactive approach addresses potential weaknesses before they impact production, ensuring maximum return on investment from CD pipeline technology. Chaos experiments can be easily added to deployment stages in Harness CD pipelines, allowing developers to verify code against various chaos scenarios to maintain and improve resilience scores. These experiments help in validating deployments against existing and new resilience conditions, platform changes, production incidents, and configuration changes, thereby preventing potential outages. The process involves creating chaos experiments, integrating them into pipelines as steps, and selecting appropriate failure strategies to manage outcomes effectively. Harness Chaos Engineering provides tools and features like chaos hubs and security governance to facilitate the seamless introduction of resilience tests into deployment stages, promoting a Continuous Resilience approach.

Harness enhances Kubernetes deployments with its continuous verification feature, integrating seamlessly with monitoring tools like Prometheus to ensure performance criteria are met and anomalies are detected early. This process supports automatic rollbacks and boosts deployment confidence, vital as software organizations focus on automation and efficiency amid heightened emphasis on security, logging, and monitoring. Continuous Verification (CV) is essential in the Continuous Delivery (CD) process, ensuring software deployments maintain functionality and performance without negatively impacting systems. Harness enables users to set up verification pipelines, utilizing unsupervised machine learning to detect anomalies, set thresholds, and automatically roll back if necessary. The platform supports various deployment strategies, such as Canary, to mitigate risks by gradually rolling out new versions, leveraging Prometheus metrics for performance validation and automated decision-making. Users can connect different logging and monitoring tools, ensuring comprehensive and reliable deployment verifications, which provides an edge in maintaining service reliability and quality.