November 2025 Summaries

Harness Database DevOps emerged from the need to make database delivery as seamless and reliable as application delivery, aiming to integrate developers, DBAs, and DevOps into a single automated workflow. The platform was developed through collaboration, research, and feedback, focusing on addressing the disconnect between application and database delivery cycles, which were traditionally handled in silos with manual processes. By leveraging existing tools like Liquibase and integrating them into a CI/CD framework, Harness sought to offer features like schema migration automation, policy compliance, and audit trails while maintaining flexibility to accommodate diverse organizational workflows. The product's development emphasized empathy-driven engineering, ensuring that database changes could be automated, visible, and governed without compromising speed or reliability, thus fostering trust between developers and data teams. The platform's evolution involved extensive testing and feedback from participants, which refined its capabilities to include features such as migration state tracking and rollback mechanisms, ultimately contributing to a more modernized approach to database management within the DevOps ecosystem.

The Shai-Hulud 2.0 incident illustrates the rapid spread of supply chain attacks when NPM maintainer accounts are compromised, affecting over 25,000 GitHub repositories and nearly 1,000 packages within hours. This attack highlights the necessity for real-time Software Bill of Materials (SBOM) visibility and policy enforcement to protect modern software supply chains. Harness Supply Chain Security (SCS) offers comprehensive solutions by providing end-to-end SBOM visibility, enabling policy enforcement to block compromised NPM packages, and ensuring complete traceability to detect malicious components early. Harness SCS allows users to identify risky components, generate Open Policy Agent (OPA) policies to block suspicious dependencies, and trace the lineage of artifacts to maintain secure pipelines. The platform facilitates swift incident response, automatic identification of vulnerabilities, and seamless integration with tools like Jira for ongoing tracking, enabling teams to effectively manage and remediate issues. By implementing these safeguards, organizations can mitigate risks, reduce exposure, and enhance the integrity of their software supply chain against attacks like Shai-Hulud 2.0.

Harness, an AWS partner, is showcasing its capabilities at AWS re:Invent 2025 by hosting an AI Survival Arena inspired by the Squid Game, where participants can engage in challenges, experience live AI demos, and explore AI-powered delivery solutions on AWS. The event, located at Booth 731 in The Venetian, Las Vegas, highlights the integration of AI in simplifying CI, CD, cost, and security processes, addressing common issues like tool sprawl in software engineering. Attendees can choose roles like Builder or Security Guardian to guide their experience, participate in strategy discussions with product leaders, and enjoy various activities, including side quests and swag giveaways. Harness also offers leadership events, such as an executive roundtable and a VIP networking dinner, to facilitate high-level discussions on AI, velocity, and risk. Dewan Ahmed, a Principal Developer Advocate at Harness, brings his extensive experience in DevOps and infrastructure to the event, emphasizing the company's commitment to enabling efficient software delivery and fostering a supportive tech community.

Harness, following its merger with Split.io, has been enhancing its platform to provide a unified and integrated experience for feature flags and experimentation, focusing on AI-driven workloads and enterprise-grade controls. The recent developments include a unified UI for Split customers, AI Experiment Summaries, rule-based segments, and Warehouse Native Experimentation, which is currently in beta. These enhancements aim to streamline processes, improve transparency, and maintain consistency across the platform without requiring code changes from developers. Harness has also been recognized as a leader in Forrester's 2025 Wave for Continuous Delivery & Release Automation, thanks to its comprehensive CI/CD and FME platform approach. The integration offers benefits like consistent user experiences, Harness-native RBAC and SSO, and SDK fallback treatments, ultimately addressing complex challenges faced by DevOps teams. As it continues to evolve, Harness plans to further integrate Split's capabilities, providing even more value to its customers.

On November 18, 2025, a major outage at Cloudflare, a key player in internet infrastructure, disrupted services globally, affecting platforms like X (formerly Twitter) and ChatGPT by causing 5xx errors and service interruptions. The outage was due to a configuration mismatch during a software rollout rather than a cyberattack, highlighting the significant risks and financial impacts of relying on third-party services. It underscored the importance of resilience testing and proactive architectural strategies to mitigate such risks, such as multi-provider redundancy, aggressive caching, robust monitoring, and graceful degradation. Harness, a company specializing in resilience solutions, advocates for regular resilience testing to identify vulnerabilities, suggesting that tools like their Chaos Engineering platform can help businesses simulate potential failures and improve their resilience posture. This incident served as a wake-up call for businesses to audit their dependencies and strengthen their systems to withstand future disruptions.

At the PASS Data Community Summit 2025, Harness will showcase advancements in Database DevOps, highlighting how automation, governance, and AI-powered tools are transforming database management to match the efficiency of application delivery. As a Bronze Sponsor, Harness will present two sessions focusing on accelerating database delivery and ensuring data integrity through AI-enabled CI/CD and automated governance. The event, held in Seattle, provides a platform for database professionals to discuss challenges and innovations, with topics such as AI-assisted development, Kubernetes-native databases, and data reliability taking center stage. Harness will demonstrate tools like AI-authored database migrations, migration tracking, automated rollback intelligence, and policy-driven governance, underscoring their commitment to modernizing and streamlining database change processes.

Harness Database DevOps has expanded its capabilities by adding Flyway support alongside Liquibase, providing teams with the flexibility to choose between structured changelogs and SQL-first migration scripts. This dual-engine approach allows developers to use their preferred tools while ensuring centralized governance, automated safety features, and a unified pipeline for database changes, thereby making database delivery as automated and reliable as application delivery. Flyway's minimalistic and convention-driven methodology appeals to developers who value direct SQL control, and its integration into Harness's platform facilitates seamless migrations with governance, drift detection, and automated rollbacks. By supporting both Flyway and Liquibase, Harness enhances developer freedom and operational confidence, promoting a tool-neutral environment that aligns with various workflows. This strategy underscores Harness's commitment to modern database delivery, offering a comprehensive pipeline experience that merges automation, policy control, and observability.

The text explores the integration of chaos engineering practices into infrastructure management using Terraform, focusing on automating various aspects to streamline resilience testing across environments. By employing the Harness Terraform provider, users can automate Kubernetes infrastructure setup, manage Git-based ChaosHub repositories, and enforce security governance policies, ensuring safe and consistent execution of chaos experiments. The document emphasizes the benefits of automation, such as reducing manual errors and setup time, by leveraging Infrastructure as Code (IaC) principles. It guides users through setting up configurations for organization and project management, Kubernetes connectors, service discovery, and custom image registries, while also addressing the implementation of security governance to control experiment execution. The guide encourages starting with non-production environments and scaling gradually, highlighting best practices for managing Terraform configurations, such as using separate state files, versioning chaos experiments, and integrating with CI/CD pipelines for automated deployment.

KubeCon NA 2025 marked a significant evolution in the cloud-native community as it embraced the "AI Native" era, with the CNCF unveiling the Kubernetes AI Conformance Program to standardize AI infrastructure. Over 9,000 attendees explored open-source innovations and discussed how AI is transforming software delivery. Harness showcased Agentic AI, which is revolutionizing software pipelines by making them intelligent and adaptive. OpenTofu Day highlighted openness, modernization, and trust, with Harness playing a key role in discussions on infrastructure delivery. Keynotes emphasized the shift from Cloud Native to AI Native, with calls for community collaboration to build this future. Security emerged as a major theme, with practical advice on securing the supply chain and the introduction of new frameworks like Apple Containerization. Technical sessions reinforced trends toward intelligent, automated, and developer-centric platforms. Harness's presence at the event underscored its commitment to leading the AI-driven future of software delivery.

The State of AI-Native Application Security 2025 report, based on a survey of 500 security practitioners from the US, UK, Germany, and France, highlights the rapid integration of AI in software development and the resulting security challenges. In 2025, AI-generated code and workflows have become integral, significantly increasing the volume of code and, consequently, the potential for vulnerabilities. While AI hasn't inherently made coding less secure, the sheer amount of code creates more opportunities for security issues, especially since traditional security tools can't keep pace with the constantly evolving AI models. A significant concern is the lack of visibility into AI's deployment within organizations, with 63% of practitioners unaware of where AI is being utilized, highlighting a gap in governance and oversight. The report emphasizes the need for security to be integrated into the development pipeline, advocating for continuous testing and validation to keep up with AI's speed. It suggests that developers require supportive measures, such as automated security checks, to innovate safely without slowing down. The integration of AI presents an opportunity to address existing chaos by improving visibility and governance throughout the workflow, ultimately aligning security and development towards shared objectives.

Harness has modernized its Commitment Orchestrator to help FinOps teams manage cloud commitments more effectively, providing a unified view of Reserved Instances, Savings Plans, on-demand, and Spot usage with enhanced metrics and streamlined workflows. This update introduces a redesigned interface, faster performance, and the foundation for AI-powered insights, allowing teams to optimize spending with less manual effort. The platform offers a centralized dashboard for visibility, integrates Spot Instances into commitment analysis, and supports detailed reporting through Net Amortized and Unblended Cost views. By employing an Atomization strategy, Harness breaks long-term commitments into smaller, manageable units to reduce lock-in risk and enhance flexibility. With an emphasis on AI-driven automation, Harness aims to transform complex cloud cost management into an intuitive process, offering features that predict, explain, and optimize spending without constant manual oversight.

Harness is tackling the AI Velocity Paradox, where rapid code generation meets bottlenecks in deployment and troubleshooting, by integrating AI deeper into the software delivery lifecycle. The company has introduced updates to its AI platform that enhance pipeline understanding, conversational troubleshooting, chat memory, and intelligent automation, thereby reducing manual toil and improving deployment stability. These innovations allow teams to comprehend complex pipelines quickly, resolve deployment failures through natural language interactions, and maintain context over time, which helps improve the continuity and efficiency of troubleshooting sessions. Harness has also enhanced its Continuous Delivery capabilities with AI-driven verification and rollback features that automatically respond to anomalies, ensuring faster, safer, and more resilient deployments. This approach aims to close the gap between the speed of code generation and the automation needed downstream, turning potential chaos into structured and intelligent software delivery.

Harness Feature Management & Experimentation (FME) has been architected to ensure continuous and resilient feature delivery, even during cloud outages, by employing a combination of automatic fallback mechanisms, distributed decision engines, and a globally distributed streaming architecture. Recent industry cloud outages highlighted the importance of reliable architecture in maintaining uninterrupted service, and Harness FME demonstrated 100% flag-delivery uptime during these disruptions. Key features of the system include graceful degradation, where SDKs automatically switch to polling without manual intervention, and resilient client-side decisioning that maintains real-time flag evaluations even during network disruptions. The separation of the control plane and delivery plane further protects user experiences by ensuring that feature delivery remains unaffected by UI performance issues. This architecture not only helps avoid disruptions but also provides a competitive advantage by ensuring reliability in unpredictable conditions, allowing for continuous improvement and learning from such events. Nico Zelaya, with extensive experience in software development and leadership, plays a pivotal role in shaping these robust systems at Harness, focusing on enhancing the developer experience and maintaining high standards of software quality and reliability.

KubeCon 2025 in Atlanta is set to focus on the integration of AI into cloud-native technologies, with a particular emphasis on improving CI/CD pipelines. Harness is showcasing its AI-powered solutions aimed at enhancing software delivery, partnering with Google Cloud to present a talk on the practical application of Agentic AI in CI/CD processes. As a key participant in the event, Harness is hosting several activities, including demos of their Agentic AI and discussions on building Internal Developer Platforms and Infrastructure as Code. The company is also a major contributor to the open-source project OpenTofu, highlighting its commitment to advancing open-source infrastructure solutions. Dewan Ahmed, a Principal Developer Advocate at Harness, brings his extensive experience in DevOps and infrastructure to the forefront, advocating for open-source and providing community support through his work and personal initiatives.

OpenTofu has emerged as a promising open-source successor to Terraform, offering a transparent and community-driven alternative following Terraform's shift to a Business Source License (BSL) in 2023. This change has led to a community-driven fork that maintains compatibility with Terraform's HCL syntax and workflows, but with enhanced features like OCI registries and encryption at rest, allowing users to manage infrastructure securely and efficiently. OpenTofu's development is guided by contributors and public RFC processes, ensuring a more agile and user-focused innovation than the centralized governance of Terraform. Harness Infrastructure as Code Management (IaCM) supports OpenTofu, enabling seamless migration and integration into existing workflows without compromising on security or scalability. The community-centric approach of OpenTofu encourages participation, allowing users to influence the future of infrastructure automation while maintaining the open-source ethos of collaboration and transparency.

Extracting a stateful service from a monolithic architecture involves carefully separating reads and writes, creating formal interfaces, and synchronizing data handling to facilitate a controlled migration to an external service. The process requires strategies such as using facades, update duplicators, and feature flag toggles to manage the transition and maintain system integrity. The article illustrates the complexity of this task with the example of extracting a Product Catalog from an e-commerce system, emphasizing the challenge of replacing database-level integrations with formal service integrations. A key strategy is to segregate querying and updating states, allowing for a phased migration while maintaining data consistency across both the monolith and the new service. The narrative underscores the critical importance of minimizing time in transitional states to prevent data inconsistency and potential corruption, advocating for precise planning and execution to ensure a successful architectural shift.

Harness Chaos Engineering can now be seamlessly integrated into Google's GKE Autopilot, allowing users to run resilience tests without compromising security. Previously, security restrictions in GKE Autopilot posed challenges for executing chaos engineering experiments, but a collaboration between Harness and Google has resolved this issue by adding Harness to an official allowlist. This integration enables chaos experiments, such as resource stress and network chaos, to be conducted within Autopilot's security boundaries without requiring any workarounds. Users can easily set up the integration by applying an allowlist and configuring Harness to operate in Autopilot mode. This development allows organizations utilizing GKE Autopilot to implement chaos engineering practices without sacrificing the benefits of a managed Kubernetes environment, thus facilitating a better understanding of application behavior under stress and enhancing system resilience.

Harness has been recognized for the second consecutive year on Inc.'s 2025 Power Partners list, which honors companies for their impactful customer relationships rather than just technological innovation. This recognition highlights Harness's dedication to collaboration, trust, and real-world outcomes, enabling clients to improve software delivery speed, security, and efficiency. By working closely with customers like Keller Williams, Ulta, and Deluxe, Harness has managed to significantly reduce test creation time and manual tasks while enhancing deployment speed. The platform's evolution is driven by customer feedback, fostering a partnership where users are key collaborators in shaping its development. Arpit Patel, a seasoned engineering leader at Harness, emphasizes the importance of this partnership in transforming technology and driving software delivery excellence, as demonstrated by successful collaborations with major companies like United Airlines and Citibank.

AI is generating more code than ever, but the delivery process hasn't kept pace, leading to inefficiencies in software development. The Harness IDP Knowledge Agent aims to bridge this gap by transforming internal developer portals (IDPs) into intelligent platforms that facilitate faster and safer software delivery. While traditional IDPs centralize service information, the Harness Knowledge Agent enhances this by acting as an AI-powered teammate that helps developers find information, understand systems, and take action swiftly. Research indicates that although AI coding assistants increase individual productivity, they do not significantly improve organizational throughput due to delivery bottlenecks and manual processes. The Knowledge Agent addresses these issues by utilizing a software delivery knowledge graph to provide contextual insights and automate tasks, making it a trusted partner in the development process. By integrating deeply into the Harness platform, it ensures data accuracy and security while offering transparency in its recommendations. This evolution from static dashboards to intelligent platforms allows developers to complete tasks more efficiently and helps organizations achieve better governance and faster release cadences, thereby aligning AI's capabilities with real-world engineering challenges.

Harness has been granted a U.S. patent for its innovative unified pipeline editor, which combines a code-based interface with a visual interface, aimed at simplifying the configuration of CI/CD pipelines. This hybrid YAML editing tool addresses the challenges posed by YAML's complexity, such as steep learning curves and high error rates, by offering both a traditional code editor enhanced with features like autocomplete and a visual editor that allows users to configure pipelines through graphical interactions. This approach not only facilitates faster onboarding and reduces errors but also broadens accessibility across various roles, from developers to non-technical stakeholders. The patented schema stitching ensures consistency and real-time validation, marking a strategic shift in software delivery by enhancing intuitive, resilient, and scalable DevOps configurations. This innovation, recognized as a unique invention, underscores Harness's commitment to improving developer experiences and accelerating software delivery in the era of microservices and platform engineering.

Traditional UI testing methods often struggle with maintenance and flakiness due to their reliance on fixed selectors, which break easily with UI changes, and their focus on implementation details rather than user outcomes. Harness AI Test Automation addresses these challenges by introducing intent-driven natural language assertions, which shift focus from how tests are written to what needs to be verified. This approach leverages AI to interpret the context and intent of tests, enabling more resilient and user-centric validation. By dynamically adapting to changes in UI and business logic, it reduces maintenance efforts, minimizes false alarms, and democratizes testing, allowing even non-technical team members to contribute effectively. Harness AI's paradigm shift in testing not only accelerates test creation and reduces maintenance but also aligns testing with real-world user experiences, thereby supporting modern DevOps practices and enhancing release velocity.