Skip to main content

Building a pipeline with Salesforce Pardot Reader

You can read data from Salesforce Marketing Cloud Engagement (formerly Pardot) using the Salesforce Pardot Reader and write to any target supported by Striim.

We recommend that you create a single Striim application using the Reader in Automated mode. In this mode, the Reader performs an initial load of existing data and automatically transitions to continuous incremental replication after the initial load has completed. For foundational concepts about pipeline phases, see Pipelines.

Reader modes

The Salesforce Pardot Reader supports three modes.

Mode

Description

When to use

Wizard

Flow Designer, TQL, API

Automated (recommended)

Performs an initial load of the selected objects and then automatically transitions to continuous incremental replication without requiring manual intervention.

Use this mode for most pipelines, especially when you are building a new pipeline.

Yes

Yes

Initial Load

Performs a one-time snapshot of the selected objects and does not continue with continuous incremental replication after completion.

Use this mode when you need to run the initial load separately from incremental replication or to perform an ad hoc resynchronization of the source data.

No

Yes

Incremental Load

Continuously replicates new or updated records using a configured watermark field and polling interval.

Use this mode if the initial load has already been completed and you want to replicate only new or updated data going forward.

No

Yes

Two ways to build a Salesforce Pardot pipeline

Before creating a pipeline, complete the required setup and configuration steps as described in Initial setup for Salesforce Pardot Reader.

You can build pipelines in one of two ways: using the pipeline wizard (recommended), or manually using Flow Designer, TQL, or the REST API.

Using the pipeline wizard (recommended)

If you are building near-real-time pipelines from Salesforce Marketing Cloud Engagement to supported targets, we recommend using a wizard from the Create App page to build an application that reads from Salesforce Pardot. The wizard automatically:

  1. Creates a single application using the Reader in Automated mode.

  2. Configures both initial load and continuous incremental replication.

  3. Creates the target schema and tables that match the selected objects.

  4. Copies existing data during the initial load phase.

  5. Automatically transitions to incremental replication when the initial load completes.

  6. Configures watermark tracking and polling intervals.

The wizard simplifies setup, promotes best practices, and reduces configuration errors.

Creating a pipeline manually

If your use case or organizational policies do not allow the use of the pipeline wizard (for example, you are using a CI/CD system to move configurations to production) or if the wizards do not support your target, you can manually create and configure an application using Flow Designer, TQL, or the Striim REST API.

When creating a pipeline manually, set the Reader's Mode property to Automated. This allows a single application to handle both initial load and continuous replication without requiring changes between phases. You may also use the Reader in Initial Load or Incremental Load mode when your architecture requires separate applications for each phase.

Configuration guidelines

When configuring the Salesforce Pardot Reader, follow these best practices.

  1. Set the Reader mode. Choose the mode that aligns with your use case. We recommend that you start with Automated mode.

  2. Identify the Incremental Load Marker. Before performing the initial load, identify a reliable, monotonically increasing field for each object. In the Incremental Load Marker property, specify this using the format <object_name>=<field_name>. If all objects share the same marker field, you can simply specify <field_name>. This field is used as the watermark for incremental replication. If an object does not have such a field, the system performs a full table copy during each incremental sync. Not all Pardot objects support incremental load; see the supported objects table in the programmer's reference for details.

  3. Configure schema creation. If the required schemas do not already exist on the target and you want Striim to create them automatically, enable Migrate Schema. If any tables from the source schema already exist on the target, this operation will fail. If you prefer not to have Striim create target schemas, pre-create target tables using native or third-party utilities and ensure that the target schema is compatible with the source schema. Leave Migrate Schema disabled in this case.

  4. Configure continuous replication. Set the Incremental Load Marker field and the Polling Interval. The polling interval must be at least 120 seconds to avoid data loss due to API inconsistency. Configure the target for upsert semantics using appropriate key columns.

  5. Set Start Timestamp (optional). If you need to start reading from a specific point in time, provide a timestamp value in the format yyyy-MM-dd HH:mm:ss.

Sample TQL for Salesforce Pardot Reader

The following TQL performs an initial load (since the default Mode is InitialLoad). For most pipelines, set Mode to Automated to perform initial load and then automatically transition to incremental replication.

CREATE SOURCE PardotIL USING SalesforcePardotReader ( 
  autoAuthTokenRenewal: true, 
  OAuthAuthorizationFlows: 'PASSWORD', 
  ThreadPoolCount: '5',
  securityToken: '<user_security_token>',
  UserName: '<username>',
  BusinessUnitId: '<pardot_business_unit_id>',
  consumerKey: '<connected_app_consumer_key>',
  consumerSecret: '<connected_app_consumer_secret>',
  SObjects: '%',
  Password: '<user_password>'
) 
OUTPUT TO PardotIL_OutputStream;