> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dqlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Salesforce Data Cloud — What We Collect

> The metadata, observability signals, lineage, and quality results Prizm extracts from Salesforce Data Cloud.

Prizm reads from the Data Cloud Metadata API and the Data Cloud query API. It never writes to Salesforce.

## Catalog Metadata

| Data Cloud object | Prizm asset | What Prizm stores |
| :- | :- | :- |
| Data Lake Object (DLO) | Table asset | Name, fields, category, and the Data Stream it belongs to |
| Data Model Object (DMO) | View asset | Name, fields, category (Profile, Engagement, or Other), and relationships to other DMOs |
| Calculated Insight | Asset | Dimensions and measures as sub-attributes |

Data Spaces map to domains in Prizm. A DMO's category is applied as a semantic tag, and key qualifier fields mark primary keys. Descriptions come from each object's label and description metadata. Data Source Objects (DSOs) appear only as the origin of a DLO in lineage, not as catalog assets.

## Observability Signals

| Signal | What it shows |
| :- | :- |
| Volume | Row count, from a `COUNT` query |
| Schema | Fields added, dropped, or changed in type between catalog runs |
| Freshness | Shown only where Data Cloud exposes a refresh timestamp through its API. Otherwise **Unknown** |
| Data Action and Data Transform alerts | Ingested as alerts when available |

## Lineage

Prizm builds native lineage from Data Cloud metadata:

* Data Source Object to DLO to DMO to Calculated Insight to Segment to Activation
* Data Transform source to target

## Quality and Criticality

* **Quality rules.** Prizm uses reusable rule templates. Key qualifier fields automatically activate uniqueness rules, and referential-integrity rules are suggested from DMO relationships. You can edit or disable any of them.
* **Exceptions.** Rows that fail a rule are found with SQL and written to a partitioned Iceberg table, linked to the rule execution.
* **Criticality.** Each object is scored from its dependencies, volume, metadata and sensitivity, key attributes, and structure. Freshness and usage aren't part of the score yet.

## What Prizm Does Not Collect

Reports and Dashboards aren't available yet. Prizm doesn't collect query usage, Data Shares, Data Graphs, CRM Analytics pipelines, or Einstein model execution. Identity Resolution rulesets, Data Streams, Data Actions, and Einstein Studio models are configuration metadata, and Prizm doesn't run or retrain them.

## Next Steps

> **[Setup](/sources/salesforce-data-cloud/setup)**
>
> Create the Salesforce app and connect Data Cloud.

> **[FAQ](/sources/salesforce-data-cloud/faq)**
>
> Common questions about the Salesforce Data Cloud connector.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.