Skip to main content

Concepts

Infrahub​

Infrahub is an inventory platform that stores information about infrastructure components such as servers, services, environments, or configurations. It acts as a centralized source of truth for discovering and querying infrastructure relationships and attributes, enabling use cases like documentation, automation, and auditing. We mainly use Vidra to automate the creation of Kubernetes manifests based on the data stored in Infrahub.

Vidra Operator​

The Vidra Operator is a Kubernetes Operator responsible for keeping Kubernetes resources in sync with inventory data and manifests stored in Infrahub. It automates the creation and update of Kubernetes native resources and Custom Resources (CRs) based on artifacts retrieved from Infrahub queries. Vidra enables continuous delivery into the Kubernetes control plane.

Artifact​

An Artifact in this context refers to a structured result (usually JSON or YAML) retrieved from Infrahub using a query. These artifacts may describe infrastructure entities such as virtual machines, web services, or any other Kubernetes resources, and are the source data Vidra uses to drive continuous delivery of Kubernetes resources.

InfrahubSync​

An InfrahubSync is a Kubernetes Custom Resource that defines a desired sync operation between Infrahub and Kubernetes. It contains:

  • A source section including:
    • The Infrahub API URL to connect to.
    • The target branch and target date to specify which version of the artifact to use.
    • An artifact name to select the correct data from Infrahub.
  • A destination section specifying:
    • The Kubernetes Server URL (optional, defaults to the current cluster).
    • The Kubernetes namespace where the resources should be created.
    • A flag to enable reconciliation on events (optional, defaults to false).
  • A status section showing the current state of the sync operation.

The Vidra Operator watches these resources and ensures the corresponding VidraResources are created or updated as needed.

VidraResource​

A VidraResource is a Kubernetes CR created and managed by the Vidra Operator based on the results of an InfrahubSync. Each VidraResource corresponds to a specific artifact from Infrahub and contains:

  • The manifest field containing the content of the artifact, representing the actual resource definition (usually in JSON or YAML format).
  • The destination section, as in InfrahubSync, specifying where the resource should be applied.
  • A status section showing the reconciliation status.

These CRs act as structured mirrors of the infrastructure state described in Infrahub.

Reconciliation​

Vidra’s reconciliation process runs when:

  • An InfrahubSync is created or updated.
  • The scheduled resync interval is triggered.
  • A VidraResource is created or updated.
  • An event occurs on a managed resource (if reconciliation on events is enabled). This is similar to auto-heal.

During reconciliation, Vidra authenticates with Infrahub, fetches the specified artifact, parses it, and ensures Kubernetes resources match the current state on the destination cluster.

Managed Resources​

Managed resources are the Kubernetes resources created or updated by Vidra based on the VidraResource manifests downloaded from Infrahub. These resources can include any Kubernetes object type, such as:

  • Deployments
  • Services
  • Namespaces
  • ConfigMaps
  • Custom Resources like VirtualMachine, etc.

Code Structure​

Vidra separates concerns through a clean architecture structure, where domain logic is kept independent of Kubernetes and Infrahub-specific implementation details.

The domain model includes types like:

  • Artifact

The Manifest itself is not saved in the domain model, but rather in the VidraResource CR. The domain model focuses on the core logic and operations related to artifacts.

This promotes testability and modularity.

To better understand the underlying technologies used in Vidra, refer to the following resources: