Resource Configurations

Place all resource definitions that Nagi needs in resources/.

All resources share the same structure as Kubernetes. The resource type is specified in the kind field.

yaml apiVersion: nagi.io/v1alpha1 kind: <resource type> metadata: name: <unique name> labels: team: data-eng annotations: description: "..." owner: "data-team@example.com" spec: ...

Attributes

Attribute	Type	Required	Default	Description
`name`	string	Yes	-	Unique name of the resource.
`annotations`	map[string, string]	—	-	Non-identifying metadata for descriptions, owner contacts, or other arbitrary information.
`labels`	map[string, string]	—	-	Key-value pairs for filtering with `--select label:key` or `--select label:key=value`.

labels

labels are used with --select selectors to filter Assets.

yaml metadata: name: daily-sales labels: team: data-eng env: prod

```bash

Select Assets that have the team label

nagi evaluate --select "label:team"

Select Assets where team=data-eng

nagi evaluate --select "label:team=data-eng" ```

Assets auto-generated by Origin (dbt) have dbt tags converted to labels with the dbt/ prefix.

yaml metadata: name: my-project.daily-sales labels: dbt/finance: "" dbt/daily: ""

annotations

annotations store supplementary information about a resource. Nagi does not use these values in its processing.

yaml metadata: name: daily-sales annotations: description: "Aggregated metrics for daily sales" owner: "data-team@example.com"

kind

Each kind has a different role within the reconciliation loop.

kind	Role in the reconciliation loop
Asset	The target of Evaluate and Sync. `onDrift` defines pairs of desired state and the corresponding convergence operation. `upstreams` references upstream Assets. `connection` specifies an external data source when desired state evaluation needs to query one
Connection	Connection information for an external data source. Used by Conditions that query data, such as `Freshness` and `SQL`
Sync	Sync procedure definition. Three stages: pre, run, post
Conditions	A resource that groups desired state definitions. Can be shared across multiple Assets
Origin	Auto-generates Assets from data structure information managed by other software
Identity	Declares an authentication scope. Referenced from Connection, Sync, and Command conditions to control which credentials are used

Template Variables

The following template variables can be used within args of Sync and Conditions. They are expanded at compile time.

Variable	Description
`{{ asset.name }}`	The Asset's `metadata.name`. For Origin-generated Assets this is prefixed with the Origin name (e.g. `my-project.orders`)
`{{ asset.modelName }}`	The original model name without the Origin prefix (e.g. `orders`). For user-defined Assets, same as `{{ asset.name }}`. Use this for external tool arguments such as `dbt run --select`
`{{ sync.<key> }}`	The value of `<key>` from the Asset's `onDrift[].with` map. For example, `{{ sync.selector }}` expands to the value of `with.selector`

`{{ asset.name }}` and `{{ asset.modelName }}`

These variables are always available and do not require with.

yaml kind: Sync metadata: name: my-sync spec: run: type: Command args: ["echo", "{{ asset.name }}"] # For Origin "my-project", model "orders": expands to "my-project.orders"

`{{ sync.<key> }}`

The Asset's onDrift[].with passes key-value pairs to the Sync. Each key becomes a {{ sync.<key> }} variable.

```yaml

Sync definition: uses {{ sync.selector }} as a placeholder

kind: Sync metadata: name: dbt-default spec: run: type: Command args: ["dbt", "run", "--select", "{{ sync.selector }}"] ```