Atlas Data Products
Turn tables into governed, discoverable data products. Atlas Data Products lets you bundle dbt models, metrics, and exposures into curated products with ownership, SLAs, and maturity tracking, all defined in Git.
Overview
Data Products live in the left navigation under Data Products. Each product is defined in a *.dataproduct.yml file committed to your repository (the same Git-native approach as Atlas Metrics). The catalog page shows a searchable grid of all products with stats, filters, and a detail sheet for deep inspection.
Defining a Data Product
A data product is a YAML file that bundles existing project assets into a governed package:
data_product:
name: revenue_analytics
title: Revenue Analytics
description: Core revenue metrics for finance reporting
owner: data-team
domain: Finance
maturity: production
access: internal
sla:
tier: gold
freshness: 1h
contains:
metrics:
- total_revenue
- monthly_recurring_revenue
- churn_rate
models:
- fct_orders
- dim_customers
exposures:
- finance_dashboard
links:
- label: Finance Dashboard
url: https://dashboard.example.com/finance
- label: Runbook
url: https://wiki.example.com/revenue-runbook
The product itself carries no metric math. It points at metrics defined in Atlas Metrics, so there is one definition governed once.
Catalog Grid
The Data Products page displays:
- Stats bar with totals for products, governed metrics, and bundled models
- Search to find products by name, domain, or description
- Maturity filter to show only production, beta, draft, or deprecated products
- Product cards showing title, description, governance badges, and asset counts
Click any card to open the detail sheet.
Governance Metadata
Each product carries governance attributes that help consumers understand trust and expectations:
Maturity
| Level | Meaning |
|---|---|
| Production | Stable, tested, and safe for downstream consumers |
| Beta | Feature-complete but still being validated |
| Draft | Work in progress, not ready for consumption |
| Deprecated | Scheduled for removal, consumers should migrate |
Access Level
| Level | Meaning |
|---|---|
| Internal | Available to your organization |
| Public | Shareable outside your organization |
| Restricted | Limited to specific teams or roles |
SLA Tiers
| Tier | Meaning |
|---|---|
| Gold | Highest freshness guarantee, mission-critical data |
| Silver | Standard freshness, suitable for most reporting |
| Bronze | Best-effort freshness, suitable for exploration |
Each tier can specify a freshness interval (e.g., 1h, 6h, 24h) representing how often the underlying data should be refreshed.
Detail Sheet
Click a product card to open the detail sheet, which shows:
- Hero header with title, domain, owner, and governance badges
- Contains section: all bundled metrics, models, dimensions, and exposures with resolution status
- Resolution health: each reference is checked against the live metric catalog and project manifest. Unresolved references (a metric that no longer exists, a model that was renamed) are flagged with a warning
- Consumer links: dashboards, documentation, and runbook URLs
Reference Validation
Data products are automatically cross-referenced against your live project:
- Metrics are checked against the Atlas Metrics catalog
- Models are checked against the dbt manifest
- Unresolved references surface as warnings in the detail sheet
- A product cannot silently claim a metric that does not exist
This keeps the catalog honest as the project evolves.
Sharing Data Products
Data products can be shared externally using Share Links:
- Click Share on any product card
- Choose between sharing the full catalog or a single product
- Generate a tokenized link with an optional expiry
- Recipients can view the product details without logging in
How It Fits Together
- Atlas Metrics: Data products reference governed metrics, so definitions stay consistent
- Slimmer CI: Breaking metric changes that affect a data product are caught in PR validation
- Atlas AI: Atlas understands data product structure and can answer questions about product contents
- Atlas Catalog: Browse documentation for the models that data products reference in the Atlas Catalog
- Scheduling: Scheduled builds keep the underlying models fresh per the product SLA
How It Compares
| Capability | dbdeux Data Products | dbt Exposures | Data catalogs (Atlan, Collibra) |
|---|---|---|---|
| Definition format | Git-native YAML in your dbt project | dbt exposure YAML | Web UI forms |
| Governance metadata | Maturity, access level, SLA tiers, domain, and owner | Owner and description only | Varies by vendor |
| Reference validation | Automatic cross-check against live metrics and manifest | No validation | Manual tagging |
| Sharing | Tokenized links with no login required | Not available | Requires platform license |
| AI integration | Atlas AI understands product structure and contents | Not available | Varies |
| Cost | Included in platform | Included in dbt Cloud | Significant additional licensing |