Skip to main content

​
Setting up a new data product

  1. Navigate to Products and click Create product
  2. Name your data product (e.g., marketing KPIs) and set a priority (i.e., P-1 for business-critical data)
  3. Assign a folder to the data product. This is useful for grouping related data products (optional)
  4. Use the Assets selection to find relevant assets. Use the browser to select or search for specific assets or annotations to select assets based on code definitions.
title

​
Setting ownership of a data product

The Data Product owner is a person or team responsible for the data product. Owners can see an overview of everything they are responsible for and understand who the owners of critical data are downstream.
  1. Choose an owner or Create a new owner to assign ownership to the data product. When you create an owner, you also set how owners are notified of issues (e.g., which Slack channel)
title

​
Setting alerts on a data product

  1. Toggle Setup alerts to active alerts for issues on the data product. Owners of the data product are notified based on the Slack or Email settings you’ve defined for the owner (see Ownership overview)
title
  • Severity—choose which type of issues alerts will include (e.g., check warn to be notified of warn-level test issues from dbt, SQLMesh, or Coalesce Transform)
  • Upstream issues—check if you want to receive alerts about upstream issues that could be impacting your data product
  • Ongoing issues—get notified about ongoing issues to ensure they get resolved (will result in more alerts)

​
Creating a data product programmatically

The asset selection you build in the wizard is a membership query, and it can be written directly rather than clicked.
  • Public API — synq.dataproducts.v2.DataproductsService lists products, fetches one with its membership definition, resolves the assets currently in it, and upserts or deletes it. Reading needs a token with SCOPE_DATAPRODUCT_READ, writing needs SCOPE_DATAPRODUCT_EDIT. See the API reference and API scopes.
  • MCP — the data product tools let an AI assistant review what exists, draft a membership query, and save the product once you approve. Reading is covered by read-only consent; saving and deleting need the Governance permission and are confirmation-gated.