Running a network means knowing how everything connects, managing who can change it, and keeping automation reliable as your environment grows. Infrahub 1.10 introduces new ways to explore network dependencies, simplify identity and access, make automation more predictable, and bring agents into how you build with and operate on Infrahub data.
Understand network dependencies
Modern networks are built from layers of interconnected systems. Routers depend on interfaces, interfaces depend on circuits, and services often span multiple parts of the network.
Understanding those relationships is essential during incident response, maintenance planning, and troubleshooting. Until now, answering even simple questions about them meant clicking through pages, writing nested GraphQL queries, or maintaining custom scripts for impact analysis.
1.10 introduces Graph Traversal, so you can trace relationships across your network and understand how systems connect without writing custom queries or maintaining separate tooling.
Graph Traversal works in two modes — Path Traversal and Dependency Discovery — each answering a different question.
How are these two objects connected?
Use Path traversal mode to understand how two objects are connected. Select a source and destination object and Infrahub will show every path between them as an interactive graph, making it easier to trace services, circuits, and dependencies across multiple layers of the network.
What will be affected if this object changes or fails?
Use dependency mode to find everything that depends on an object. Starting from a router, circuit, or service, you can identify all connected objects of a given type and quickly understand the impact of a failure or Proposed Change.
Available wherever you work
You can use Graph Traversal from:
- The Infrahub UI
- GraphQL
- The Python SDK
- The Infrahub MCP Server
In the UI, every object detail page has a Trace from this object action that opens the explorer pre-seeded with that object.
Graph Traversal always reflects the context you’re working in. If you’re operating on a branch, the results reflect the state of that branch — or a chosen point in time — and existing permissions continue to determine what data is visible.
It’s also available through the MCP Server, so agents can trace the same relationships your engineers rely on every day.
You can investigate dependencies and run impact analysis directly in Infrahub. For example:
- What services will be affected if this router goes down?
- What systems depend on this circuit?
- What is the impact of this Proposed Change?
Simplify identity and access management
As more people begin using Infrahub, onboarding users and managing access often becomes additional administrative work. Infrahub 1.10 introduces several improvements that align user management with the identity systems you already operate.
Authenticate with your existing directory services (Infrahub Enterprise)
You can now authenticate directly against any LDAP server using native LDAP support on Infrahub Enterprise, alongside existing local account and SSO options. User accounts are automatically created the first time someone signs in.
Reduce manual group management
Instead of pre-creating groups, Infrahub can automatically build them from claims provided by your identity provider or LDAP directory, based on rules you define.
You can onboard users faster, reduce manual administration, and maintain a cleaner audit trail without changing your existing roles and permissions.
Learn more:
Build smarter automation
The cost of a change should match the size of the change. Updating a single object should not trigger unnecessary repository-wide rebuilds or automation runs.
These updates reduce unnecessary artifact generation, make it easier to understand why work was triggered, and help keep large repositories fast and reliable as the amount of data and change activity grows.
Regenerate only what changed
When you commit a change to a connected Git repository, Infrahub now rebuilds only the artifacts affected by that change — or artifacts whose definitions have changed. Generators now run only when a change touches fields they actually use.
Speed up large merges
Merge operations now occur directly at the database layer, making large merges faster and reducing the likelihood of difficult recovery scenarios.
Preview migrations before applying them
These are the core schema migrations that run during an upgrade, and they are easier to manage in 1.10. infrahub db migrate can now preview the migration plan without applying changes, while infrahub db showmigrations reports what has already run and what remains pending.
Infrahub Sync 2.0
The latest release of Infrahub Sync introduces several improvements for synchronizing data between Infrahub and other systems. Rather than re-reading an entire source on every execution, Sync now processes only changed records and loads them in parallel, reducing execution time for scheduled synchronizations. Sync can also derive object creation order directly from schema mappings, removing the need to manually maintain dependency ordering as your data model evolves.
Additional improvements include:
- Record-count checks that stop a run when a source shrinks unexpectedly, guarding against accidental mass deletions
- Saved diffs and change plans for later review
- Faster recurring synchronization jobs
These improvements help keep automation predictable as data, repositories, and teams grow.
Learn more: Infrahub Sync 2.0 release notes
Build with agents and Infrahub
Many teams are beginning to incorporate agents into how they build and operate network automation. Infrahub 1.10, together with the latest releases of the MCP Server and AI Skills, gives you practical ways to bring those tools into your existing workflows.
The MCP Server and AI Skills address two different needs. The MCP Server gives agents access to Infrahub data, while AI Skills provide task-specific guidance and best practices for building with Infrahub.
AI Skills: Build with Infrahub more effectively
You can use Infrahub AI Skills — each designed for a specific task — to create schemas, generators, transforms, validation checks, integrations, and user interface components while automatically applying Infrahub conventions and best practices. You can also use the skills to investigate dependencies, detect drift, audit data quality, and analyze live Infrahub data.
When you’re tackling larger initiatives, Specification-Driven Development (SDD) lets you review and refine a plan before any resources are created.
Recent improvements include:
- Branch-first workflows by default for changes created through the CLI, MCP tools, generators, and SDK
- More efficient generators that can reuse query responses instead of repeatedly fetching related objects
- A diagnostics collector that generates redacted support bundles
- A skill that files sanitized bug reports in the appropriate repository
- Support for file-bearing nodes through CoreFileObject
MCP Server: Use Infrahub data throughout your environment
Deploy the MCP Server so agents can work with Infrahub data and relationships outside of Infrahub itself. Instead of reconstructing context from multiple systems, agents can use the same objects, relationships, and branch context your engineers already rely on every day.
Recent improvements include:
- Deploying the server centrally through Docker Compose or Helm
- Preserving user permissions and audit trails through pass-through authentication
- Read-only mode and health endpoints for production deployments
- Create and update objects in a dedicated branch and open a single Proposed Change for everything you build in one conversation
You can now use agents to create schemas, build generators, investigate dependencies, analyze Proposed Changes, and use Infrahub data in other systems.
Learn more:
Developer experience improvements
Automation works best when applications can understand exactly what went wrong and how to respond.
Machine-readable GraphQL errors
Every GraphQL error now includes a stable error code, a typed payload, and an HTTP status, allowing clients to branch on machine-readable responses instead of parsing text. When multiple fields fail validation, every error is returned in a single response, so forms and applications can highlight all issues at once.
More control when designing schemas
Schema authors gain more control over ordering, including support for metadata fields such as created_at and updated_at. The same ordering can now be applied to the results of GraphQL queries at request time, using the order argument. Reserved names and inheritance conflicts are also surfaced earlier during schema development.
Learn more:
Before you upgrade
Infrahub 1.10 includes several breaking changes. Review the release notes before upgrading, particularly if you:
- Parse GraphQL error responses
- Use custom schemas that inherit from internal Core generics
- Use node_metadata
- Use BuiltinIPPrefix.resource_pool
- Manage SSO user migrations
More resources
- Introducing Infrahub Skills. Learn what the skills are and how to start using them in your AI tool.
- Infrahub Skills Livestream Recap. Get a walkthrough of the skills in action, with examples from the live session.
