June 2026 Summaries

OpsMill is now the steward of the Nornir project, a Python-native infrastructure orchestration framework that has become essential for many infrastructure teams over the past eight years. Emphasizing its commitment to keeping Nornir open-source and community-driven, OpsMill plans to modernize and support the project while fostering collaboration with the community to shape its future. Founded by David Barroso, who created Nornir and previously developed NAPALM, the project benefits from his vision of a framework that simplifies network automation without introducing new complexities. OpsMill values Barroso's ongoing role as a Technical Advisor, and many of its team members have longstanding involvement with Nornir, contributing to its success and supporting the open-source community. The company invites feedback from users and contributors to better understand their needs and explore potential commercial support offerings, demonstrating its dedication to ensuring Nornir's continued growth and relevance in the evolving landscape of infrastructure automation.

OpsMill has become the steward of Nornir, the Python-native open-source infrastructure automation framework created by David Barroso in 2017, and says it will preserve the project’s community-driven nature while providing maintenance, resources, and modernization efforts. The company views Nornir as an important execution layer for emerging AI and autonomous infrastructure systems, which require reliable data, clear operational controls, and predictable, observable automation. Rather than immediately publishing a roadmap, OpsMill plans to develop future priorities with users, contributors, and plugin developers. Barroso, who also created NAPALM, will remain connected to OpsMill as a technical advisor but will not manage Nornir’s daily stewardship. OpsMill highlights its longstanding involvement in the Nornir ecosystem, including maintainers and plugin contributors on its team, and invites community feedback while exploring potential demand for enterprise-grade support.

Infrahub 1.10 brings significant enhancements to network management by introducing Graph Traversal, which allows users to explore network dependencies without custom queries, and improving identity and access management with native LDAP support for streamlined user onboarding. The update also optimizes automation by ensuring only necessary artifacts are regenerated after changes, thus maintaining fast and reliable repositories. Additionally, Infrahub Sync 2.0 increases synchronization efficiency by processing only changed records, while the integration of agents and AI Skills into workflows aids in schema creation, dependency analysis, and data auditing. The MCP Server facilitates the use of Infrahub data across environments, and developer experience is enhanced through machine-readable GraphQL errors and improved schema design controls. The release requires attention to breaking changes documented in its release notes.

Infrahub 1.10 adds Graph Traversal for visualizing paths between network objects and identifying dependencies or potential change impacts through the UI, GraphQL, Python SDK, and MCP Server, while respecting branch context, historical state, and user permissions. The release expands enterprise identity management with native LDAP authentication, automatic account creation, and rule-based group assignment from directory or identity-provider claims. Automation improvements include selective artifact regeneration, database-layer merges, migration previews, and Infrahub Sync 2.0, which processes changed records in parallel, derives dependency order from schemas, and adds safeguards such as record-count checks and saved change plans. New AI Skills and MCP Server capabilities support agent-assisted schema design, automation development, data analysis, dependency investigation, and controlled branch-based updates, with centralized deployment, pass-through authentication, audit trails, and read-only modes. Developer-focused changes provide machine-readable GraphQL errors, improved schema and query ordering controls, and earlier detection of naming and inheritance conflicts, while users are advised to review breaking changes involving error parsing, custom schemas, metadata, IP prefix resources, and SSO migrations before upgrading.

Infrahub 1.10 adds Graph Traversal capabilities for visualizing paths between network objects and identifying dependencies or potential change impacts through the UI, GraphQL, Python SDK, and MCP Server, while respecting branch context and existing permissions. The release also expands enterprise identity management with native LDAP authentication, automatic user creation, and directory-driven group assignment, and improves automation by limiting artifact regeneration to affected changes, accelerating database-layer merges, and enabling migration previews. Infrahub Sync 2.0 processes changed records in parallel, infers dependency order from schemas, and adds safeguards such as record-count checks and saved diffs. Agent-oriented updates to Infrahub Skills and the MCP Server support guided development, dependency investigation, centralized deployment, permission pass-through, audit trails, and branch-based changes. Additional developer improvements include structured GraphQL errors, enhanced schema ordering controls, earlier validation of naming and inheritance conflicts, and upgrade guidance for several breaking changes.

AIOps, or Artificial Intelligence for IT Operations, is an appealing concept where AI agents autonomously manage and optimize network operations, yet its practical implementation faces significant challenges. Current AI agents often malfunction unless tasked with highly specific duties, causing skepticism and reluctance to deploy them in production environments. This mirrors past experiences with network automation, where rushed implementations led to fragmented solutions due to poor foundational planning. A successful approach involves modeling intent, structuring data, and prioritizing automation only after establishing a robust foundation. The Network Automation Framework (NAF) offers a blueprint for scalable, reliable network automation, emphasizing six foundational elements: intent, executor, observability, collector, orchestrator, and presentation. These elements support a structured, tool-agnostic, and continuously updated framework that is accessible to various stakeholders, not just network engineers. The NAF Framework provides a systematic approach to network automation, enabling seamless integration of AI while ensuring reliability and scalability through thorough planning and execution.

The narrative highlights the challenges faced by organizations trying to integrate AI agents into their operations, particularly when dealing with outdated and fragmented data models. It draws attention to the common issues such as the lengthy process of updating schemas, reliance on complex and often makeshift solutions, and the lack of a unified source of truth which results in inefficiencies and errors. The text emphasizes the necessity for a comprehensive, flexible, and well-maintained data model that provides both technical and business context to enable AI agents to function effectively. Examples from industries like telecommunications and global companies such as TikTok and Eurofiber illustrate the problems of misaligned data models, the need for human intervention, and the importance of having a data model that accurately reflects the current state and purpose of the infrastructure to avoid costly mistakes and to fully leverage automation capabilities.

AI agents in network infrastructure require accurate, current, and context-rich data models to operate safely, yet many organizations face fragmented sources of truth, slow schema changes, improvised toolchains, and disconnected automation templates. Examples from companies including TikTok, ByteDance, Eurofiber, and financial-services and telecom teams illustrate how rigid or poorly owned data systems can delay updates, force engineers into manual spreadsheets and custom integrations, and leave parts of the network undocumented. These conditions can cause agents to act on incomplete or outdated information, potentially repeating configuration errors, overlooking dependencies, or making decisions without awareness of ownership, costs, capacity, compliance, or business purpose. The proposed remedy is a flexible but structured shared data foundation that represents both the intended technical state of the network and the operational context behind it before agentic automation is deployed.

OpsMill has developed a decentralized documentation system where each team is responsible for maintaining the documentation alongside the code in their respective repositories, addressing the common issue of outdated documentation due to unclear ownership. This approach utilizes a push-based sync mechanism, identified as Option C, where changes in documentation automatically update a central Docusaurus site, ensuring that the documentation remains current and aligned with the code changes. The system avoids the pitfalls of a monolithic documentation repository and cumbersome git submodule management by allowing each project's documentation to be self-contained while contributing to a unified platform. The setup includes a multi-repo architecture with individual plugin instances for each project within the Docusaurus site, enabling seamless integration and search functionality across projects. This method enhances the consistency and accuracy of the documentation, benefiting customers with up-to-date information and streamlining processes for engineers and new joiners. OpsMill emphasizes the importance of detailed configuration and scripting to maintain consistency across multiple repositories, with the goal of facilitating efficient onboarding and information dissemination.

Infrahub is positioned as an automation-focused infrastructure platform that diverges from the traditional asset inventory approach by prioritizing a context-heavy source of truth for network management. Unlike asset management systems that track ownership, cost, and lifecycle of hardware for financial and procurement purposes, Infrahub emphasizes understanding the intended and actual state of a network to facilitate reliable automation and AI applications. This approach allows for effective configuration generation, service provisioning, and drift reconciliation by integrating device and warranty data into its system, ensuring that finance, procurement, and network automation can collaborate effectively. Infrahub achieves this integration through Infrahub-sync, which imports essential data from IT Asset Management (ITAM) or Configuration Management Database (CMDB) systems to enable informed decision-making for network changes while maintaining a clear distinction between asset tracking and network modeling.

OpsMill argues that asset management and infrastructure automation require distinct but connected systems: asset inventories focus on financial and procurement information such as serial numbers, purchase orders, warranties, invoices, and lifecycle status, while automation-oriented platforms model the intended and actual operational state of networks, including device-service relationships and their history. It presents Infrahub as an automation-first source of truth designed to support version-controlled network modeling, configuration generation, service provisioning, drift reconciliation, change validation, and rollback workflows. The company says its Infrahub Sync integration brings selected asset or CMDB data, including device, serial number, and warranty details, into the automation model so organizations can account for hardware lifecycle constraints when planning or deploying network changes.

OpsMill and its platform, Infrahub, are pioneering a shift in the infrastructure management landscape by emphasizing a version-controlled, policy-aware system of record that integrates governance and flexibility. As the industry gravitates toward these features, OpsMill's approach validates the necessity of a robust infrastructure source of truth, especially in the context of AI-driven network automation. Infrahub's key features include a flexible, user-defined schema that adapts to dynamic infrastructure needs, native version control for safe and auditable changes, and built-in governance to streamline validation and deployment processes. This infrastructure model is designed to accommodate evolving technologies and ensure seamless integration between technical data and business logic, addressing challenges posed by AI data centers' complexity and rapid advancements. As companies evaluate their data management platforms, they are encouraged to consider solutions like Infrahub that are purpose-built for this transition, promising to redefine capabilities over the coming years.

OpsMill argues that recent additions of branching, pull-request-style governance, and procurement workflows by established infrastructure platforms reflect an industry shift toward version-controlled, policy-aware infrastructure systems of record. It contends that effective automation and AI require accurate, centralized infrastructure data rather than fragmented or outdated information, and presents Infrahub as built around three core capabilities: user-defined, code-based schemas that model diverse and changing infrastructure environments; native version control through a temporal graph database that supports historical records, branching, merging, review, rollback, and parallel work; and integrated governance for validation, artifact generation, approvals, and impact review. The piece links this direction to growing organizational pressure to adopt AI and to the complexity of modern AI data centers, whose networks involve large GPU clusters, specialized topologies, and interconnected device, optical, cooling, and power relationships. It concludes that platform architecture will increasingly determine whether teams can safely connect infrastructure intent, operational changes, automation pipelines, and AI agents over the coming years.