December 2025 Summaries

Dynamic Pipelines in Harness provide a flexible CI/CD solution by allowing pipeline configurations to be generated and executed at runtime, rather than being pre-defined and stored. This approach addresses the limitations of traditional "configuration as code" models, especially in scenarios requiring adaptable workflows, such as AI-driven decisions, frictionless migrations, and programmatic orchestration. Dynamic Pipelines support "headless" orchestration and facilitate migrations by converting existing pipeline definitions into Harness YAML. Despite their complexity, they offer significant advantages for certain applications, ensuring governance through role-based access control, OPA policies, and secure management of secrets. The feature is API-driven, enabling immediate execution of pipelines by passing YAML payloads or integrating dynamic stages within stable pipeline frameworks. By enabling Dynamic Execution at the account and pipeline levels, teams can leverage this advanced functionality to create "paved path" CI/CD patterns without sacrificing security or compliance.

Harness's India organization has experienced significant growth, expanding to 480 employees and aiming for over 1,000 within three years, driven by a focus on increasing technical depth and impact, particularly in AI and R&D. This growth aligns with the software industry's shift towards addressing post-code complexities such as testing, security, and cost management, emphasizing the need for unified intelligence and experienced teams. Harness's Bangalore AI R&D Center contributes to various areas, including machine learning, platform architecture, and security intelligence, and plays a crucial role in developing and operating critical parts of the platform. The company is committed to building diverse talent pipelines and engaging with the local developer community to support long-term ecosystem development. With recent funding, Harness continues to invest in its platform and teams, particularly in India, to enhance software testing, security, and deployment as AI integration increases, thereby focusing on solving complex problems and building scalable systems.

AI is transforming software delivery by emphasizing the need for robust fundamentals rather than replacing them, according to Nick Durkin, Field CTO at Harness. By 2026, teams will rely on specialized AI agents for specific tasks, moving away from the misconception of a singular, omniscient AI system. This shift will necessitate the replacement of traditional governance models with flexible guardrails, allowing for speed and innovation while maintaining security and compliance. The integration of DevSecOps will become a reality as security is embedded within the development lifecycle, enabling immediate feedback and reducing friction between teams. Engineers will evolve into managers of AI agents, focusing on context and creative problem-solving rather than mundane coding tasks. Ultimately, AI will not solve existing software delivery issues but will highlight them, allowing well-structured teams to thrive while others may struggle with existing inefficiencies.

Vedant Shrotria discusses the development of a developer-friendly end-to-end (E2E) testing framework for chaos fault validation aimed at reducing setup friction while maintaining control and correctness in chaos engineering. Previously, the process to validate chaos faults was cumbersome, requiring manual installation of dependencies, configuration of environment variables, and YAML-based workflows, which hindered feedback loops and adoption. The new framework offers an API-driven model, real-time log streaming, and intelligent target discovery, alongside dual-phase validation to ensure both fault impact and recovery are verified. The architecture includes an Experiment Runner for orchestrating the experiment lifecycle, an Experiment Monitor for status tracking, and a Validation Framework for concrete chaos impact verification, allowing for faster execution and scalability, with test setups now taking under five minutes. The framework is proprietary but highlights best practices that can be applied to similar testing infrastructures, emphasizing the importance of developer experience, automation, and knowledge sharing to enhance testing processes.

Knowledge graphs and Retrieval-Augmented Generation (RAG) are complementary techniques that enhance large language models with external knowledge, particularly useful for DevOps. A knowledge graph is a semantic model that maps entities and relationships within systems, ensuring consistent definitions and enabling multi-hop reasoning, while RAG retrieves unstructured text based on semantic similarity, excelling in documentation search and open-ended queries. The hybrid approach combines the structured reasoning of knowledge graphs with the contextual breadth of RAG, creating a robust framework for DevOps automation by providing structured context and unstructured information. This synergy, supported by a semantic layer, allows for tasks like context-aware pipeline generation and graph-grounded debugging, resulting in more reliable and efficient DevOps processes. Harness's implementation exemplifies this by integrating a Software Delivery Knowledge Graph with RAG, leading to significant improvements in pipeline onboarding speed, issue resolution, and debugging efficiency, demonstrating the benefits of combining these methodologies for a more comprehensive DevOps intelligence system.

Cloud migration has evolved from a mere tactical exercise of moving applications between clouds to a comprehensive strategic program that encompasses infrastructure redesign, governance enhancement, and cost management. Enterprises face challenges in this endeavor due to complex systems, compliance requirements, parallel operations, and varied CI/CD tools. Automation, as provided by platforms like Harness, is crucial in mitigating risks, ensuring consistency, and reducing manual efforts during the migration process. This structured approach not only aids in compliance and cost control but also modernizes delivery practices and strengthens governance, which is particularly vital in regulated sectors. As organizations transition, they often pursue cloud migration to improve performance, access better-managed services, or support a multi-cloud strategy, with a focus on creating predictable, auditable, and secure processes. The series introduces these high-level concepts and will delve deeper into technical architectures and financial operations in subsequent entries.

Harness Database DevOps now supports Google AlloyDB, providing a comprehensive platform for managing PostgreSQL-compatible schema changes with CI/CD, GitOps, and policy-driven governance, which enhances database delivery by reducing operational risk and manual overhead. The integration with Google Cloud's AlloyDB, a high-performance, scalable, and PostgreSQL-compatible database engine, is crucial for mission-critical applications in cloud modernization efforts. Harness enables seamless migration and operationalization of AlloyDB through automated workflows, ensuring consistent, predictable, and auditable database changes using Liquibase or Flyway. This integration helps organizations achieve a 99% reduction in manual schema deployment overhead and enables end-to-end CI/CD automation, improving resilience and governance. The synergy between Harness and AlloyDB promotes secure, scalable, and modern database delivery aligned with DevOps principles, enhancing the speed and reliability of software deployments.

Jyoti Bansal, co-founder of Harness, discusses the company's mission to automate and streamline software delivery processes, emphasizing that the real bottleneck in software engineering occurs after code is written, in the stages of testing, verification, security, deployment, and governance. With the rise of AI, the volume of code production has increased, exacerbating the challenges faced by manual workflows and fragmented tools. Harness has raised $240 million in a Series E financing round, which values the company at $5.5 billion, to further develop its AI-driven platform aimed at managing the complexities of software delivery. Bansal highlights Harness's capabilities, including its knowledge graph, AI agents, and orchestration engine, which help organizations accelerate their delivery workflows by reducing manual effort and increasing efficiency. He notes the significant improvements seen by customers like United Airlines, Morningstar, Keller Williams, National Australia Bank, and Citibank, and underscores the importance of delivering software safely and reliably in the AI era.

Harness Workflows, part of the Internal Developer Portal (IDP), aim to transform traditional ticket-driven operations by providing developers with a self-service model that accelerates software delivery while maintaining consistency and governance. These Workflows automate and streamline the process of creating and managing services, environments, and operational tasks, reducing the downtime developers face due to manual and often fragmented processes. By encapsulating best practices, security guidelines, and compliance rules within automated pathways, Workflows allow developers to focus on coding rather than navigating complex infrastructure and administrative tasks, ultimately enhancing productivity and sustainability. This approach not only empowers developers but also enables platform engineering teams to scale their expertise efficiently, supporting a more agile and responsive software development lifecycle.

Developer velocity and DBA caution are not opposing forces but rather two essential priorities that can coexist through the implementation of Database DevOps, which introduces automated validation, approvals, and visibility to facilitate collaboration and trust. Platforms like Harness enable developers and DBAs to work together by integrating schema changes into code pipelines, offering a shared context that enhances empathy through automation, thus transforming potential conflicts into partnerships. This approach mitigates the historical friction caused by different priorities, as developers focus on speed while DBAs prioritize data integrity and performance. By aligning process, empathy, and automation, the divide between developers and DBAs disappears, resulting in a shared mission to deliver reliable software. This harmony is achieved not by replacing DBAs but by empowering them with tools that allow them to focus on strategic tasks, ensuring that both precision and speed are maintained in software delivery.

A critical unauthenticated Remote Code Execution (RCE) vulnerability, CVE-2025-55182, has been identified in React Server Components and Next.js, presenting a severe threat with a CVSS score of 10.0. This vulnerability, discovered by Lachlan Davidson, affects the "Flight" protocol in React's server-side rendering, allowing unauthorized attackers to execute arbitrary code by exploiting insecure deserialization. Although the vulnerability impacts numerous frameworks, Traceable by Harness WAF provided immediate protection against this class of vulnerabilities through multi-layered defenses such as Server Side Template Injection and Node.js Injection attack rules. The vulnerability affects specific versions of React and Next.js and requires immediate attention from organizations using these technologies to ensure protection through Traceable WAF and timely patching. The disclosure underscores the importance of proactive security measures and research-driven innovation to guard against evolving threats, with Traceable emphasizing rapid deployment of defenses and continuous improvement to stay ahead of potential exploits.

AI has significantly accelerated software delivery processes, participating in every stage of the delivery pipeline, yet it has also introduced new security challenges that outpace traditional safety measures. This dynamic shift has exposed gaps between the speed of delivery and the assurance of safety, as AI-driven automation can amplify existing weaknesses, alter threat landscapes, and create automated hazards due to misconfigurations. Developers often lack adequate training to address AI-specific security issues. At the DevSecOps 2025 conference, Dewan Ahmed proposed a four-pillar framework to address these challenges, emphasizing contextual intelligence, automatic verification, behavior-based anomaly detection, and continuous learning loops to ensure secure and trustworthy AI-enhanced delivery systems. Dewan Ahmed, a seasoned developer advocate and speaker, has extensive experience in solving DevOps and infrastructure problems and is committed to fostering a secure and inclusive tech community.

Harness AI is advancing its capabilities by enhancing its partnership with Amazon to integrate AI-powered development with intelligent delivery, focusing on secure, efficient code deployment in production. This collaboration aims to streamline workflows using AI-infused tools, improving processes like pipeline onboarding and debugging, while maintaining governance and security across AWS environments. Harness is also addressing database management challenges by introducing AI-powered database migration authoring, which enables developers to use natural language to generate production-ready, policy-compliant migrations, enhancing DevOps speed and governance. The platform's Error Analyzer has been improved to quickly identify and resolve pipeline errors, offering actionable insights and automated fixes, thereby reducing debugging time and maintaining reliability. Furthermore, Harness AI continually optimizes its model stack, ensuring it selects the best AI models for specific tasks, and maintaining cutting-edge performance in software delivery. These developments, alongside continuous model evaluation and AWS integration, are part of Harness's broader vision to create an AI-native software delivery ecosystem that enhances code velocity while ensuring robust pipeline, database, and platform performance.

Harness Infrastructure as Code Management (IaCM) introduces Variable Sets and Provider Registry as tools to enhance the efficiency and security of Terraform and OpenTofu workflows. These innovations address challenges such as configuration drift, secret duplication, and secure distribution of custom providers, which are common when scaling Infrastructure as Code (IaC) across numerous workspaces and teams. Variable Sets offer a centralized control plane for managing configuration parameters and secrets, ensuring consistency and reducing manual efforts during credential updates. Provider Registry provides a trusted mechanism for distributing custom providers, incorporating GPG-signed binaries to ensure security and integrity across platforms. Together, these features enable platform teams to streamline infrastructure management, enforce governance, and facilitate collaborations without hindering developer productivity, extending Harness IaCM's capabilities as a comprehensive, AI-driven infrastructure management solution.

Database DevOps and Database Migration Systems address different aspects of database workflows, where the former focuses on collaboration, governance, and automation, while the latter ensures structured and versioned schema execution. These systems work together to streamline database delivery by integrating seamlessly into CI/CD processes, thereby alleviating the common bottleneck of database changes in automated workflows. Harness exemplifies this integration by combining the cultural framework of DevOps with the structured approach of migration systems, facilitating seamless, safe, and compliant database changes. Harness supports tools like Liquibase OSS and Flyway, enabling developers to commit schema updates, have DBAs review them, and automatically apply migrations through secure pipelines, complete with audit trails and real-time rollbacks. This synergy allows teams to treat databases as dynamic components of their products, enhancing collaboration, trust, and the ability to scale efficiently, thereby transforming database delivery into a streamlined extension of CI/CD pipelines.

Harness, a company specializing in software delivery solutions, has announced an expanded partnership with Amazon, integrating Amazon Kiro, Amazon Q Developer, and Harness SaaS on AWS. This collaboration aims to enhance software development and deployment by leveraging AWS's infrastructure to provide a scalable, intelligent, and seamless delivery experience for AWS customers. Harness SaaS on AWS offers features such as accelerated deployments, AI-driven automation, continuous verification, and automated security and compliance, enabling faster and safer software releases. Additionally, it provides cloud cost optimization and integrates with Amazon Kiro and Q Developer, allowing developers to manage CI/CD pipelines and security directly from their IDE using natural language. Companies like Trust Bank have benefited from this integration by significantly reducing deployment lead times and improving infrastructure resilience and security. Harness's AI-powered solutions, including Test Intelligence and Continuous Verification, are designed to streamline the software delivery process, helping organizations focus on innovation rather than manual processes. This expanded partnership is available through the AWS Marketplace, allowing organizations to easily adopt and scale Harness's software delivery solutions.

Warehouse Native Experimentation allows organizations to conduct experiments directly within their data warehouses, like Snowflake and Amazon Redshift, eliminating the need for data exports to external systems. This approach provides teams with faster, more trustworthy, and transparent results because experiments are conducted using the same trusted data sources the business relies on. By integrating experimentation with internal data models and governance controls, teams can make data-driven decisions with greater confidence. This methodology supports the creation and reuse of metrics that reflect business goals, ensuring that experiments align with organizational objectives. As product velocity becomes a competitive advantage, this approach helps maintain speed and compliance, reducing risks associated with launching new features. With Warehouse Native Experimentation, teams benefit from full visibility into experiment data, allowing for comprehensive analysis, validation, and collaboration across departments, from product development to data science. Looking forward, the integration of these workflows into CI/CD pipelines aims to further streamline the experimentation process, enhancing both efficiency and confidence in product releases.