Databricks source overview

What this is

Databricks is a unified analytics platform built around Apache Spark. Zeotap CDP integrates with it as a data warehouse source under the Integrate module, pulling records from a Databricks table on the schedule you choose so your CDP profiles stay in sync. This page is the entry point for the Databricks source: it explains the two supported connection modes, what you gather from Databricks before creating the source, how to verify the source is working, and where to find the full field-by-field setup.

Databricks works as both a source and a destination. This page covers the source side — pulling data from Databricks into Zeotap CDP. If you want to activate audiences into Databricks tables instead, see Databricks destination.

Prerequisites

Before you create a Databricks source in Zeotap CDP, collect the following from your Databricks workspace:

Databricks Host — the workspace URL, taken from your browser address bar on the Databricks account.
Catalog Name, Schema Name, and Table Name — the three-level identifier that points at the table you want to ingest.
Connection credentials — either OAuth for JDBC (Client ID + Client Secret + HTTP Path) or a personal access token with a Cluster ID and Warehouse ID for the Job-based mode.

The field-by-field walkthrough — including how to create the service principal, generate the OAuth secret, and locate the HTTP Path — is documented in Create and Implement a Databricks Source.

Keep credentials out of shared channels. The Client Secret, the personal access token, and any service-account JSON grant access to your Databricks workspace and the linked cloud-storage bucket. Do not paste them into tickets, chat threads, or version control.

Choose a connection mode

Zeotap CDP supports two mechanisms for pulling data from Databricks. Choose based on the size of the table you are ingesting.

Mode	When to use	Credentials required
JDBC	Tables under 1 million records.	OAuth Client ID, Client Secret, and HTTP Path. The HTTP Path is found under SQL Warehouses → Connection details in Databricks.
Job-based	Recommended for tables above 1 million records.	Personal access token (or service-principal credentials), Cluster ID, and Warehouse ID.

For tables above 1 million records, start with Job-based. The setup guide flags that JDBC can hit performance issues at that volume.

What you configure in the source form

The Zeotap CDP source form has three settings you decide up front, in addition to the Databricks connection fields:

Sync Frequency — how often Zeotap CDP re-reads the Databricks table. Supported cadences are Sync once, Every hour, Every 3 hours, Every 6 hours, Every 12 hours, Daily, Weekly, and Monthly. For Daily / Weekly / Monthly you also pick a Sync Time (AM/PM) and, for Monthly, the day of the month.
Data Entity — mark the source as Customer Data or Non Customer Data, depending on whether the records are customer profiles or reference / lookup data.
Delta Queries Selection — set to true to fetch only rows that changed since the last run (based on a timestamp column you name), or false to re-ingest the full table on every run.

The step-by-step form entry is documented in Set Up a Databricks Source. Values you collect from Databricks map to the source form as follows:

# Fields to enter in the Zeotap CDP Databricks source form

# Source details (both modes)
databricks_host:   # Workspace URL from your browser address bar
catalog_name:      # Unity Catalog / Metastore catalog holding the schema
schema_name:       # Database inside the catalog
table_name:        # Target table to ingest

# JDBC mode credentials
client_id:         # OAuth client ID from the Databricks service principal
client_secret:     # OAuth secret for that service principal
http_path:         # Databricks -> SQL Warehouses -> Connection details

# Job-based mode credentials
access_token:      # Databricks -> Settings -> User Settings -> Developer -> Generate new token
cluster_id:        # Databricks -> Compute -> cluster name -> automatically added tags
warehouse_id:      # Databricks -> SQL Warehouses -> select warehouse -> Warehouse ID

Verify the source is working

You know the source is configured correctly when:

The Databricks source appears in the Sources listing with a status of Integrated.
The Implementation Details tab on the source shows the connection parameters you entered.
The first data sync starts once the source is created; subsequent runs follow the Sync Frequency you selected.

The initial data transfer scales with table volume — large tables take longer than small ones to complete the first sync.

Why a source fails to connect

Error string	What it means	Where to look
Source does not reach the Integrated status after creation	Zeotap CDP could not connect to Databricks with the credentials supplied.	Re-check the Databricks Host value and re-verify the OAuth client or access token in your Databricks account.
First sync starts but does not finish on a table above 1 million records	JDBC hits performance limits at large volume.	Recreate the source using the Job-based connection mode instead of JDBC.
OAuth credentials rejected in JDBC mode	The service principal is missing workspace permissions, or the OAuth secret has been rotated or revoked.	Databricks → the service principal’s workspace-level permissions and OAuth secrets.
Access token rejected in Job-based mode	The personal access token has expired or been revoked.	Databricks → Settings → User Settings → Developer → regenerate the token, then edit the source with the new value.

If the source still does not reach Integrated after re-checking the credentials and the connection mode, contact the Zeotap support team at support@zeotap.com. Include the source name, the connection mode selected (JDBC or Job-based), the Databricks Host value, and the timestamp of the failed sync.

​What this is

​Prerequisites

​Choose a connection mode

​What you configure in the source form

​Verify the source is working

​Why a source fails to connect

​Next steps

What this is

Prerequisites

Choose a connection mode

What you configure in the source form

Verify the source is working

Why a source fails to connect

Next steps