Prerequisites
Before configuring the dbt connector in Prizm, ensure the following:- dbt Cloud (API mode): A dbt Cloud account with a service token that has at least Member permissions on the target project.
- dbt Core (file mode): Access to the
manifest.jsonandrun_results.jsonfiles generated bydbt compileordbt run. - The dbt project must target a data source already connected to Prizm (e.g., Snowflake, BigQuery, Databricks).
Connection Modes
| Mode | How It Works |
|---|---|
| dbt Cloud (API) | Prizm polls the dbt Cloud API on a schedule to fetch the latest manifest, run results, and catalog artifacts from a specified job. |
| dbt Core (File Upload) | Upload manifest.json and optionally run_results.json directly to Prizm. Useful for self-hosted or CI-driven dbt Core setups. |
Connection Fields
dbt Cloud (API Mode)
| Field | Description |
|---|---|
| Connection Name | A display name for this dbt integration in Prizm (e.g., dbt-prod). |
| Account ID | Your dbt Cloud account ID, found in the dbt Cloud URL: cloud.getdbt.com/accounts/{account_id}. |
| Service Token | A dbt Cloud service token with Member or higher permissions. Generate one under Account Settings → Service Tokens. |
| Project ID | The numeric ID of the dbt Cloud project to sync. Found in the project URL. |
| Job ID | The ID of the dbt Cloud job whose artifacts Prizm will fetch. Prizm reads the most recent completed run for this job. |
| Environment | Optional. Filter by dbt Cloud environment (e.g., Production). |
dbt Core (File Upload Mode)
| Field | Description |
|---|---|
| Connection Name | A display name for this dbt integration in Prizm. |
| manifest.json | Upload the manifest.json file from your dbt project’s target/ directory. |
| run_results.json | Optional. Upload run_results.json to sync test pass/fail results into Prizm. |
What Prizm Ingests
| Artifact | What Prizm Extracts |
|---|---|
| manifest.json | Model definitions, column descriptions, model-level tests, source references, and cross-model lineage (DAG). |
| run_results.json | Test execution results (pass, fail, warn, error) mapped to Prizm data quality dimensions. |
| catalog.json*(Cloud only)* | Column-level metadata and statistics from the warehouse to enrich Prizm’s catalog. |
Asset Scope
Prizm maps each dbt model to the corresponding table or view in the connected data source. Only models whose target schema and database are covered by an existing Prizm connector are linked. Models that reference sources outside Prizm’s scope are imported as reference nodes only (lineage edges are preserved but the nodes are not profiled).Job Schedules
| Job Type | Default Frequency | What It Does |
|---|---|---|
| Catalog Sync | Daily | Fetches the latest manifest and updates model metadata, descriptions, and tags in Prizm. |
| Test Results Sync | Every hour (Cloud) / On upload (Core) | Pulls dbt test results and maps them to Prizm quality checks. |
| Lineage Rebuild | Daily | Rebuilds the end-to-end lineage graph incorporating dbt model DAG edges. |
For dbt Cloud mode, Prizm fetches artifacts from the most recently completed run of the specified job. If the job has never completed successfully, no artifacts will be available and the sync will be skipped.
What’s next
Snowflake Connector
Connect the Snowflake warehouse that your dbt project targets.
Lineage
See how Prizm merges dbt DAG edges with warehouse query lineage.
Data Quality Tests
Understand how dbt test results map to Prizm quality dimensions.
Catalog Sync
Learn how dbt model descriptions and tags enrich the Prizm catalog.
Prerequisites
Before configuring the dbt connector in Prizm, ensure the following:- dbt Cloud (API mode): A dbt Cloud account with a service token that has at least Member permissions on the target project.
- dbt Core (file mode): Access to the
manifest.jsonandrun_results.jsonfiles generated bydbt compileordbt run. - The dbt project must target a data source already connected to Prizm (e.g., Snowflake, BigQuery, Databricks).
Connection Modes
| Mode | How It Works |
|---|---|
| dbt Cloud (API) | Prizm polls the dbt Cloud API on a schedule to fetch the latest manifest, run results, and catalog artifacts from a specified job. |
| dbt Core (File Upload) | Upload manifest.json and optionally run_results.json directly to Prizm. Useful for self-hosted or CI-driven dbt Core setups. |
Connection Fields
dbt Cloud (API Mode)
| Field | Description |
|---|---|
| Connection Name | A display name for this dbt integration in Prizm (e.g., dbt-prod). |
| Account ID | Your dbt Cloud account ID, found in the dbt Cloud URL: cloud.getdbt.com/accounts/{account_id}. |
| Service Token | A dbt Cloud service token with Member or higher permissions. Generate one under Account Settings → Service Tokens. |
| Project ID | The numeric ID of the dbt Cloud project to sync. Found in the project URL. |
| Job ID | The ID of the dbt Cloud job whose artifacts Prizm will fetch. Prizm reads the most recent completed run for this job. |
| Environment | Optional. Filter by dbt Cloud environment (e.g., Production). |
dbt Core (File Upload Mode)
| Field | Description |
|---|---|
| Connection Name | A display name for this dbt integration in Prizm. |
| manifest.json | Upload the manifest.json file from your dbt project’s target/ directory. |
| run_results.json | Optional. Upload run_results.json to sync test pass/fail results into Prizm. |
What Prizm Ingests
| Artifact | What Prizm Extracts |
|---|---|
| manifest.json | Model definitions, column descriptions, model-level tests, source references, and cross-model lineage (DAG). |
| run_results.json | Test execution results (pass, fail, warn, error) mapped to Prizm data quality dimensions. |
| catalog.json*(Cloud only)* | Column-level metadata and statistics from the warehouse to enrich Prizm’s catalog. |
Asset Scope
Prizm maps each dbt model to the corresponding table or view in the connected data source. Only models whose target schema and database are covered by an existing Prizm connector are linked. Models that reference sources outside Prizm’s scope are imported as reference nodes only (lineage edges are preserved but the nodes are not profiled).Job Schedules
| Job Type | Default Frequency | What It Does |
|---|---|---|
| Catalog Sync | Daily | Fetches the latest manifest and updates model metadata, descriptions, and tags in Prizm. |
| Test Results Sync | Every hour (Cloud) / On upload (Core) | Pulls dbt test results and maps them to Prizm quality checks. |
| Lineage Rebuild | Daily | Rebuilds the end-to-end lineage graph incorporating dbt model DAG edges. |
For dbt Cloud mode, Prizm fetches artifacts from the most recently completed run of the specified job. If the job has never completed successfully, no artifacts will be available and the sync will be skipped.
What’s next
Snowflake Connector
Connect the Snowflake warehouse that your dbt project targets.
Lineage
See how Prizm merges dbt DAG edges with warehouse query lineage.
Data Quality Tests
Understand how dbt test results map to Prizm quality dimensions.
Catalog Sync
Learn how dbt model descriptions and tags enrich the Prizm catalog.