Home > Docs > Features > OneLake Catalog

📚 OneLake Catalog - Unified Data Discovery & Governance Hub¶

Last Updated: 2026-04-13 | Version: 1.0.0

📑 Table of Contents¶

• 🎯 Overview
• 🏗️ Architecture
• ⚙️ Discovery Features
• 🏷️ Endorsement & Certification
• 🛡️ Govern Tab
• 🔍 Search API & MCP
• 🎰 Casino Implementation
• 🏛️ Federal Agency Implementation
• ⚠️ Limitations
• 📚 References

🎯 Overview¶

OneLake Catalog is Microsoft Fabric's unified data discovery hub that provides a centralized experience for browsing, searching, tagging, endorsing, and governing all items across your Fabric estate. Available as a Generally Available feature, OneLake Catalog replaces the fragmented experience of navigating between workspaces, Lakehouses, Warehouses, and semantic models to find the right dataset -- instead offering a single searchable catalog that spans the entire organization's Fabric footprint.

OneLake Catalog sits at the intersection of data discovery and data governance. While data engineers create and curate assets in workspaces, OneLake Catalog makes those assets findable and trustworthy for consumers across the organization. It integrates with Microsoft Purview for sensitivity labeling and classification, supports endorsement workflows for data quality certification, and provides a governance dashboard that gives domain owners visibility into label coverage, endorsement status, and usage patterns.

Key Capabilities¶

Why OneLake Catalog Matters¶

Where OneLake Catalog Fits¶

flowchart TB
    subgraph Producers["👤 Data Producers"]
        DE["Data Engineers"]
        DS["Data Scientists"]
        BI["BI Developers"]
    end

    subgraph Catalog["📚 OneLake Catalog"]
        SEARCH["🔍 Search & Discovery"]
        ENDORSE["🏷️ Endorsement"]
        GOVERN["🛡️ Governance"]
        META["📋 Metadata Index"]
    end

    subgraph Items["💾 Fabric Items"]
        LH["🏠 Lakehouses"]
        WH["🏢 Warehouses"]
        SM["📊 Semantic Models"]
        NB["📓 Notebooks"]
        DF["🔄 Dataflows"]
        EH["⚡ Eventhouses"]
    end

    subgraph Consumers["👤 Data Consumers"]
        AN["Analysts"]
        EX["Executives"]
        APP["Applications"]
        AGT["AI Agents"]
    end

    Producers --> Items
    Items --> META
    META --> SEARCH
    META --> ENDORSE
    META --> GOVERN
    SEARCH --> Consumers
    ENDORSE --> Consumers

    style Catalog fill:#6C3483,stroke:#4A235A,color:#fff
    style Items fill:#2471A3,stroke:#1A5276,color:#fff
    style Consumers fill:#27AE60,stroke:#1E8449,color:#fff
    style Producers fill:#E67E22,stroke:#CA6F1E,color:#fff

🏗️ Architecture¶

OneLake Catalog is a metadata service that indexes all Fabric items across workspaces and domains, providing a queryable discovery layer without duplicating the underlying data. The catalog ingests metadata from each workspace, builds a search index, and exposes it through both the Fabric portal UI and a REST API.

Component Architecture¶

flowchart LR
    subgraph Workspaces["🔷 Fabric Workspaces"]
        WS1["Workspace A<br/>Casino Analytics"]
        WS2["Workspace B<br/>Federal Data"]
        WS3["Workspace C<br/>Shared Services"]
    end

    subgraph Indexing["📋 Metadata Indexing"]
        CRAWL["Item Crawler"]
        SCHEMA["Schema Extractor"]
        TAG["Tag Aggregator"]
        LINEAGE["Lineage Builder"]
    end

    subgraph CatalogCore["📚 OneLake Catalog Core"]
        IDX["Search Index"]
        ENDORSE["Endorsement<br/>Registry"]
        LABEL["Sensitivity<br/>Label Store"]
        USAGE["Usage<br/>Analytics"]
    end

    subgraph Integration["🔌 Integrations"]
        PURVIEW["Microsoft Purview<br/>Classification"]
        API["Search API<br/>(Preview)"]
        MCP["MCP Tool<br/>(Preview)"]
        PORTAL["Fabric Portal<br/>UI"]
    end

    Workspaces --> CRAWL
    CRAWL --> SCHEMA
    CRAWL --> TAG
    CRAWL --> LINEAGE
    SCHEMA --> IDX
    TAG --> IDX
    LINEAGE --> IDX
    IDX --> PORTAL
    IDX --> API
    API --> MCP
    ENDORSE --> PORTAL
    LABEL <--> PURVIEW

    style Workspaces fill:#2471A3,stroke:#1A5276,color:#fff
    style Indexing fill:#E67E22,stroke:#CA6F1E,color:#fff
    style CatalogCore fill:#6C3483,stroke:#4A235A,color:#fff
    style Integration fill:#27AE60,stroke:#1E8449,color:#fff

Metadata Flow¶

1. Item Crawler continuously monitors Fabric workspaces for new, modified, or deleted items
2. Schema Extractor reads table schemas, column names, data types, and descriptions from Lakehouses, Warehouses, and Semantic Models
3. Tag Aggregator collects user-defined tags, endorsement status, and sensitivity labels
4. Lineage Builder traces data flow relationships from source (Bronze) through transformations (Silver) to consumption (Gold/Reports)
5. Search Index consolidates all metadata into a queryable index supporting full-text search and faceted filtering
6. Endorsement Registry tracks promotion and certification status with approval audit trails
7. Usage Analytics aggregates access patterns for trending items and popularity signals

Access Control Model¶

OneLake Catalog respects Fabric workspace permissions. Users only see items in the catalog that they have at least Viewer access to in the source workspace.

💡 Tip: OneLake Catalog does not grant data access. Discovering an item in the catalog does not mean you can query its data. Users must have appropriate permissions on the underlying workspace item (Lakehouse, Warehouse, etc.) to access the actual data.

⚙️ Discovery Features¶

Full-Text Search¶

OneLake Catalog search covers all indexed metadata including item names, descriptions, tags, table names, column names, and measure definitions.

Search scope:

Search operators:

# Basic keyword search
slot telemetry

# Exact phrase
"bronze slot telemetry"

# Filter by item type
slot telemetry type:Lakehouse

# Filter by endorsement
revenue type:SemanticModel endorsed:certified

# Filter by domain
crop production domain:USDA

# Combine filters
compliance type:Lakehouse domain:Casino endorsed:promoted

Faceted Filtering¶

The catalog UI provides faceted filters on the left panel for interactive refinement:

Recently Accessed¶

The "Recently Accessed" section on the catalog home page shows the last 20 items you opened, queried, or modified. This provides a personal favorites-like experience without requiring explicit bookmarking.

Trending Items¶

"Trending Items" surfaces the most-viewed items across workspaces you have access to, based on aggregated usage signals over the past 7 days. This helps new team members discover important datasets organically.

Lineage View¶

From any catalog item, click the Lineage button to see the full upstream and downstream dependency graph. The lineage view shows:

• Upstream: Source systems, ingestion pipelines, Bronze tables, transformations
• Current item: Highlighted in the graph
• Downstream: Dependent reports, dashboards, notebooks, and derived datasets

flowchart LR
    subgraph Sources["📡 Sources"]
        EH["Event Hub<br/>Slot Telemetry"]
        API["REST API<br/>Player Data"]
    end

    subgraph Bronze["🥉 Bronze"]
        BT["bronze_slot_telemetry"]
        BP["bronze_player_activity"]
    end

    subgraph Silver["🥈 Silver"]
        ST["silver_slot_cleansed"]
        SP["silver_player_validated"]
    end

    subgraph Gold["🥇 Gold"]
        GS["gold_slot_performance"]
        GP["gold_player_kpis"]
    end

    subgraph Consumption["📊 Consumption"]
        SM["Semantic Model<br/>Casino Analytics"]
        RPT["Power BI Report<br/>Floor Dashboard"]
        NB["Notebook<br/>ML Scoring"]
    end

    Sources --> Bronze --> Silver --> Gold --> Consumption
    SM --> RPT

    style Bronze fill:#CD7F32,stroke:#8B5A2B,color:#fff
    style Silver fill:#C0C0C0,stroke:#808080,color:#000
    style Gold fill:#FFD700,stroke:#B8860B,color:#000
    style Consumption fill:#6C3483,stroke:#4A235A,color:#fff

🏷️ Endorsement & Certification¶

Endorsement is OneLake Catalog's mechanism for signaling data quality and trustworthiness. It replaces ad-hoc naming conventions (e.g., "final_v2_approved") with formal badges that are visible across the entire catalog.

Endorsement Levels¶

Promotion Workflow¶

Promotion is a lightweight endorsement that any workspace Member can apply. It signals "I recommend this item" without requiring a formal approval process.

sequenceDiagram
    participant DE as Data Engineer
    participant CAT as OneLake Catalog
    participant TEAM as Team Members

    DE->>CAT: Create Lakehouse item
    DE->>CAT: Add description + tags
    DE->>CAT: Set endorsement = "Promoted"
    CAT->>CAT: Display 🔵 Promoted badge
    TEAM->>CAT: Search for "slot telemetry"
    CAT->>TEAM: Return item with Promoted badge
    TEAM->>TEAM: Trust signal: team-recommended

    Note over DE,TEAM: Promotion is instant, no approval needed

Certification Workflow¶

Certification is a formal endorsement that requires Admin privileges and implies organizational review. It is the highest trust signal in the catalog.

sequenceDiagram
    participant DE as Data Engineer
    participant QA as Quality Assurance
    participant ADMIN as Workspace Admin
    participant CAT as OneLake Catalog
    participant ORG as Organization

    DE->>CAT: Request certification
    DE->>QA: Submit quality checklist
    QA->>QA: Validate schema, completeness, accuracy
    QA->>ADMIN: Recommend certification
    ADMIN->>CAT: Set endorsement = "Certified"
    CAT->>CAT: Display ✅ Certified badge
    CAT->>CAT: Record audit trail (who, when, why)
    ORG->>CAT: Search for certified datasets
    CAT->>ORG: Return items with Certified badge

    Note over DE,ORG: Certification implies organizational review

Endorsement Best Practices¶

Endorsement Configuration (Admin)¶

Tenant admins can control endorsement policies at the tenant and domain level:

Fabric Admin Portal > Tenant Settings > Content Certification

- Enable certification: ✅
- Certification allowed for: "Specific security groups"
- Security groups: "Data Stewards", "Domain Admins"
- Documentation link: "https://wiki.contoso.com/data-certification-policy"

🛡️ Govern Tab¶

The Govern tab provides domain owners and workspace admins with a governance dashboard that surfaces compliance gaps, endorsement coverage, and sensitivity label status across all items in their scope.

Governance Dashboard Metrics¶

Governance by Domain¶

The Govern tab supports domain-scoped views. When domains are configured (see Data Mesh patterns), domain admins see governance metrics only for items within their domain:

flowchart TB
    subgraph GovernTab["🛡️ Govern Tab"]
        OVERALL["Overall Health Score<br/>87/100"]
    end

    subgraph Domains["📁 Domain Views"]
        D1["🎰 Casino Domain<br/>Score: 92/100"]
        D2["🌾 USDA Domain<br/>Score: 85/100"]
        D3["💼 SBA Domain<br/>Score: 78/100"]
        D4["🌀 NOAA Domain<br/>Score: 90/100"]
        D5["🌊 EPA Domain<br/>Score: 88/100"]
    end

    subgraph Metrics["📊 Per-Domain Metrics"]
        M1["Endorsement Coverage"]
        M2["Label Coverage"]
        M3["Description Coverage"]
        M4["Stale Items"]
    end

    GovernTab --> Domains
    Domains --> Metrics

    style GovernTab fill:#6C3483,stroke:#4A235A,color:#fff
    style Domains fill:#2471A3,stroke:#1A5276,color:#fff
    style Metrics fill:#27AE60,stroke:#1E8449,color:#fff

Sensitivity Label Integration¶

OneLake Catalog integrates with Microsoft Purview Information Protection to display and manage sensitivity labels:

💡 Tip: Use Purview auto-labeling policies to automatically classify new Lakehouse tables based on column names (e.g., any table with an ssn or social_security_number column gets labeled "Highly Confidential" automatically).

Quality Score Calculation¶

The governance health score (0-100) is calculated as a weighted average:

🔍 Search API & MCP¶

OneLake Catalog Search API (Preview)¶

The Search API provides programmatic access to OneLake Catalog for automation, CI/CD validation, and AI agent integration. As of April 2026, the API is in Public Preview.

API Endpoints¶

# Search catalog items
POST https://api.fabric.microsoft.com/v1/catalog/search
Content-Type: application/json
Authorization: Bearer {access_token}

{
    "query": "slot telemetry",
    "filters": {
        "itemType": ["Lakehouse", "Warehouse"],
        "endorsement": ["Certified"],
        "domain": ["Casino"]
    },
    "top": 10,
    "skip": 0,
    "orderBy": "relevance"
}

Response:

{
    "value": [
        {
            "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
            "name": "lh_bronze",
            "itemType": "Lakehouse",
            "workspaceId": "ws-casino-001",
            "workspaceName": "Casino Analytics",
            "description": "Bronze layer Lakehouse for raw casino floor telemetry",
            "endorsement": "Certified",
            "sensitivityLabel": "Confidential",
            "domain": "Casino",
            "owner": "jdoe@contoso.com",
            "lastModified": "2026-04-10T14:30:00Z",
            "tags": ["bronze", "slot-telemetry", "compliance"],
            "tables": [
                {
                    "name": "bronze_slot_telemetry",
                    "columns": ["machine_id", "timestamp", "event_type", "coin_in", "coin_out"]
                }
            ]
        }
    ],
    "totalCount": 3,
    "continuationToken": null
}

Python SDK Integration¶

import requests
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()
token = credential.get_token("https://api.fabric.microsoft.com/.default")

# Search for certified Lakehouses in the Casino domain
response = requests.post(
    "https://api.fabric.microsoft.com/v1/catalog/search",
    headers={
        "Authorization": f"Bearer {token.token}",
        "Content-Type": "application/json"
    },
    json={
        "query": "slot",
        "filters": {
            "itemType": ["Lakehouse"],
            "endorsement": ["Certified"],
            "domain": ["Casino"]
        },
        "top": 10
    }
)

results = response.json()
for item in results["value"]:
    print(f"  {item['endorsement']}: {item['name']} ({item['workspaceName']})")
    print(f"    {item['description']}")

CI/CD Validation: Ensure Endorsement Before Deployment¶

# Validate that all Gold-layer items are endorsed before production deployment
def validate_endorsement(workspace_name: str, layer: str = "gold") -> bool:
    """Ensure all items matching the layer pattern are endorsed."""
    response = requests.post(
        "https://api.fabric.microsoft.com/v1/catalog/search",
        headers={"Authorization": f"Bearer {token.token}",
                 "Content-Type": "application/json"},
        json={
            "query": layer,
            "filters": {
                "workspace": [workspace_name],
                "endorsement": ["None"]
            },
            "top": 100
        }
    )

    unendorsed = response.json()["value"]
    if unendorsed:
        print(f"❌ {len(unendorsed)} {layer}-layer items are NOT endorsed:")
        for item in unendorsed:
            print(f"  - {item['name']} ({item['itemType']})")
        return False

    print(f"✅ All {layer}-layer items in {workspace_name} are endorsed")
    return True

# Use in deployment pipeline
assert validate_endorsement("Casino Analytics", "gold"), \
    "Deployment blocked: unendorsed Gold items"

MCP Tool Integration (Preview)¶

The OneLake Catalog MCP tool enables AI agents (such as Fabric Data Agents or external MCP clients) to search the catalog using natural language, discover relevant datasets, and understand data lineage programmatically.

# MCP tool definition for OneLake Catalog search
# Registered as an MCP server tool
{
    "name": "onelake_catalog_search",
    "description": "Search OneLake Catalog for Fabric items by keyword, type, domain, or endorsement status",
    "parameters": {
        "query": "string - Search keywords",
        "item_type": "string - Filter by item type (optional)",
        "domain": "string - Filter by domain (optional)",
        "endorsed_only": "boolean - Return only endorsed items (optional)"
    }
}

Agent interaction example:

User: "Find the certified slot telemetry dataset"

Agent → MCP Tool → OneLake Catalog API:
  query="slot telemetry", endorsement=["Certified"]

Agent: "I found the certified Bronze Lakehouse 'lh_bronze' in the 
Casino Analytics workspace. It contains the 'bronze_slot_telemetry' 
table with columns: machine_id, timestamp, event_type, coin_in, 
coin_out. Last updated April 10, 2026."

💡 Tip: Combining the OneLake Catalog MCP tool with Fabric Data Agents creates a powerful self-service data discovery experience. Analysts can ask "What certified datasets do we have for EPA compliance?" and the agent uses the catalog API to return relevant, trusted items.

🎰 Casino Implementation¶

Domain Setup for Gaming¶

Configure the Casino domain to group all gaming-related items in the catalog:

Fabric Admin Portal > Domains > New Domain

Name: Casino Gaming
Description: All casino floor analytics, compliance, and player management items
Owners: casino-data-stewards@contoso.com
Workspaces: Casino Analytics, Casino Compliance, Casino Real-Time

Endorsement Strategy for Gaming Items¶

Compliance Dataset Certification Checklist¶

Before certifying any compliance-related dataset (CTR, SAR, W-2G), the following checks must pass:

Casino Catalog Search Examples¶

# Find all compliance-related certified items
compliance type:Lakehouse endorsed:certified domain:Casino

# Find SAR-related notebooks
sar type:Notebook domain:Casino

# Find all items tagged with "regulatory"
tag:regulatory domain:Casino

# Find W-2G reporting datasets
w-2g type:SemanticModel endorsed:certified

Tagging Strategy¶

Apply consistent tags to improve discoverability:

🏛️ Federal Agency Implementation¶

Cross-Agency Catalog with Domain Boundaries¶

Each federal agency operates as a separate domain in OneLake Catalog, providing clear governance boundaries while enabling cross-agency discovery for authorized users.

flowchart TB
    subgraph Catalog["📚 OneLake Catalog"]
        SEARCH["Unified Search"]
    end

    subgraph Domains["📁 Agency Domains"]
        D1["🌾 USDA<br/>Agricultural Data"]
        D2["💼 SBA<br/>Loan Programs"]
        D3["🌀 NOAA<br/>Weather & Climate"]
        D4["🌊 EPA<br/>Environmental"]
        D5["🏔️ DOI<br/>Natural Resources"]
    end

    subgraph SharedItems["🔗 Cross-Agency"]
        GEO["Geospatial<br/>Reference Data"]
        CAL["Calendar<br/>Dimensions"]
        STATE["State/Region<br/>Dimensions"]
    end

    SEARCH --> Domains
    SEARCH --> SharedItems
    D1 -.->|"Shared dims"| SharedItems
    D2 -.->|"Shared dims"| SharedItems
    D3 -.->|"Shared dims"| SharedItems
    D4 -.->|"Shared dims"| SharedItems
    D5 -.->|"Shared dims"| SharedItems

    style Catalog fill:#6C3483,stroke:#4A235A,color:#fff
    style Domains fill:#2471A3,stroke:#1A5276,color:#fff
    style SharedItems fill:#27AE60,stroke:#1E8449,color:#fff

Domain Configuration per Agency¶

FISMA Labeling for Federal Data¶

Federal Information Security Modernization Act (FISMA) requires classification of all federal information systems. Map FISMA impact levels to Purview sensitivity labels:

Federal Endorsement Policy¶

Agency Data Governance Policy:

1. All Bronze-layer items: No endorsement (raw ingestion, not validated)
2. All Silver-layer items: Promoted (schema-validated, team-reviewed)
3. All Gold-layer items: Certified (quality-checked, compliance-approved)
4. Cross-agency shared dimensions: Certified (approved by data governance board)

Certification Requirements (Federal):
- Data quality score > 95% (Great Expectations suite pass)
- Sensitivity label applied
- FISMA classification documented
- Data dictionary complete
- Lineage documented from source API to Gold table
- Section 508 accessibility compliance for reports

Cross-Agency Discovery Examples¶

# Find all certified crop production data
crop production endorsed:certified domain:USDA

# Find weather data that's publicly shareable
weather type:Lakehouse label:Public domain:NOAA

# Find all loan-related items across agencies
loan type:Lakehouse

# Find EPA enforcement data with proper sensitivity
enforcement endorsed:certified label:Confidential domain:EPA

# Discover shared dimension tables
type:Lakehouse tag:shared-dimension

# Find all items modified in the last 7 days across agencies
modified:last7days type:Lakehouse

Cross-Agency Governance Dashboard¶

The federal governance dashboard aggregates metrics across all agency domains:

⚠️ Warning: Cross-agency data sharing must comply with each agency's data sharing agreements (DSAs) and Privacy Impact Assessments (PIAs). OneLake Catalog visibility does not imply authorization to access or combine data across agency boundaries. All cross-agency access must be documented and approved through the federal data governance board.

⚠️ Limitations¶

Current Limitations¶

What is Not Supported¶

⚠️ Warning: OneLake Catalog is not a replacement for Microsoft Purview Data Catalog in multi-cloud or hybrid environments. For organizations that need to catalog data assets across Azure, AWS, GCP, and on-premises systems, use Purview Data Catalog with OneLake Catalog as the Fabric-specific complement.

📚 References¶

🔗 Related Documents¶

• OneLake Security -- Fine-grained access control for items discovered in the catalog
• Data Mesh Enterprise Patterns -- Domain-based organization that drives catalog structure
• Data Agents -- AI agents that use catalog search for data discovery
• Fabric IQ -- Natural language querying that can leverage catalog metadata
• Architecture -- System architecture overview

📝 Document Metadata - Author: Documentation Team - Reviewers: Data Governance, Security, Compliance, Federal Data Stewards - Classification: Internal - Next Review: 2026-07-13