> ## 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.

# Azure Synapse Overview

> What the Azure Synapse connector covers, what Prizm collects from each compute surface, and how it authenticates.

Prizm connects to an Azure Synapse Analytics workspace to catalog its tables, views, and columns, monitor freshness, volume, and schema changes, run data quality checks, and observe Synapse pipelines. One Prizm source covers one Synapse workspace, and you choose which parts of the workspace to include. Prizm is read-only on Synapse.

## Why Connect Azure Synapse to Prizm?

**Catalog & Context** Prizm discovers the databases, schemas, tables, views, and columns in your Dedicated and Serverless SQL pools, along with distribution, index, and key details on Dedicated pools.

**Data Quality** Structural, business, lookup, behavioral, and conditional quality checks run on your own Synapse compute, and failing records are written to an Iceberg table rather than back to Synapse.

**Observability** Volume, freshness, and schema changes are tracked for every cataloged table, alongside query performance for your pools.

**Pipelines** Prizm catalogs Synapse pipelines, activities, triggers, and data flows, and monitors run status, duration, and failures.

**Lineage & Criticality** Prizm derives table-level lineage from pipeline definitions, view definitions, foreign keys, external tables, and CTAS/CETAS statements, and scores each table's criticality.

## Compute Surfaces

The connection form lets you choose which surfaces Prizm covers. Each has its own prerequisites, so grant access only for the ones you select.

| Surface | What Prizm covers |
| :- | :- |
| Dedicated SQL pool | Tables, views, columns, distribution strategy and key, index type, row counts, keys, observability, quality, usage, and lineage |
| Serverless SQL pool | External tables over Azure Data Lake Storage Gen2 (CSV, Parquet, ORC) and Delta tables exposed as external tables |
| Pipelines | Pipelines, activities, triggers, data flows, and run monitoring |

## Authentication

| Type | What you provide | Surfaces |
| :- | :- | :- |
| Service Principal (client secret) | Tenant ID, Client ID, and Client Secret | All |
| Service Principal (certificate) | Tenant ID, Client ID, and a PEM certificate with its private key | All |
| Managed Identity | Nothing for a system-assigned identity, or the Client ID of a user-assigned one | All, when Prizm runs on Azure |
| SQL Authentication (Dedicated pool only) | A SQL username and password | Dedicated SQL pool only |

Turn on **Use Vault** to retrieve credentials from a vault instead of entering them. See [Setup](/sources/synapse/setup) for the permissions each type needs.

## Current Scope

The connector is read-only. Some items are limited today:

* Cost management (DWU, bytes scanned, and similar) isn't part of the connector, and Prizm doesn't request cost permissions.
* A paused Dedicated pool is reported as paused and skipped. Prizm never resumes it.
* Lineage is table-level. Column-level lineage and deep notebook or Spark code parsing aren't covered.
* Dedicated pool freshness is approximate, because Dedicated pools have no native last-modified time. See [What We Collect](/sources/synapse/what-we-collect).

## Next Steps

> **[Setup](/sources/synapse/setup)**
>
> Grant permissions and connect your Synapse workspace.

> **[What We Collect](/sources/synapse/what-we-collect)**
>
> The metadata, signals, and lineage Prizm extracts.

> **[FAQ](/sources/synapse/faq)**
>
> Common questions about the Synapse connector.

> **[Troubleshooting](/help/troubleshooting)**
>
> Diagnose connection, permission, and performance errors.


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