August 2026

Connector SDK

fivetran-connector-sdk PyPI package

We have released version 2.11.0 of the fivetran-connector-sdk PyPI package. This release includes the following updates:

• Adds support for errors and warnings.
• Introduces the --yes flag and updates other non-interactive flags for better clarity and user experience.
• Adds resolution precedence for required flags and providing the configuration.json path.
• Updates the new connection creation and debug flows to require configuration.json. For more information, see configuration resolution order.
• Updated fivetran init and AI plugin setup behavior for existing projects.
• Minor enhancements and improvements.
• Updated testers.

fivetran-api-playground PyPI package

We have released version 1.1.9 of the fivetran-api-playground PyPI package. This release upgrades the faker library dependency to 40.36.0.

Insights Tab

We've added the Insights tab to the Connection Details page in the Fivetran dashboard. It highlights the value Fivetran provides for your connection.

You can now view key maintenance details about your connections, including:

• Schema changes handled
• Source deletes captured
• Volume of data processed
• Associated downstream models
• Recent connector improvements

Learn more in our Insights tab documentation.

Managed Data Lake Service

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the data lake, which could cause differences between the source and destination data.

Partner SDK

Partner SDK Tester

We have released version 2.26.0810.001 of the Partner SDK Tester. This release includes the following improvements:

• Added HTML rendering support for errors and warnings messages to improve readability during local testing.

Schema changes

Braze

We have added two new tables, CAMPAIGN_ANALYTICS and CAMPAIGN_ANALYTICS_VARIATION.

Canvas Data 2 by Instructure

D2L Brightspace

Fivetran Platform Connector

We've added a new column, connection_details, to the CONNECTION and GROUP tables.

GitHub

We've added a new table, ORG_MEMBER.

We have added a new column, role, to the TEAM_MEMBERSHIP table.

HubSpot

We have added the following new tables to sync data for the Courses object:

• COURSE
• COURSE_COMPANY
• COURSE_CONTACT
• COURSE_DEAL
• COURSE_ENGAGEMENT
• COURSE_PIPELINE
• COURSE_PIPELINE_STAGE
• COURSE_PROJECT
• COURSE_PROPERTY_HISTORY
• COURSE_TICKET

We are gradually rolling out this change to all connections.

LinkedIn Ad Analytics

We've added a new table, VIDEO. The table contains video media assets referenced by posts, including the download URL, its expiration time, duration, aspect ratio, thumbnail, owner, and processing status.

OpenAI

We have added the following new tables to store organization admin and usage data:

• CODE_INTERPRETER_SESSION
• FILE_SEARCH_CALL
• IMAGE
• MODERATION
• PROJECT_API_KEY
• VECTOR_STORE_USAGE
• WEB_SEARCH_CALL

Ordway

Pinterest Ads

We've added a new column, optimization_goal_metadata, to the AD_GROUP_HISTORY table.

We've added a new table, ADVERTISER_TARGETING_GEO_REPORT. The table contains account-level impressions, clicks, spend, eCPM, and engagement metrics broken down by country, region, location, and ZIP-code geo targeting.

Rocketlane

Salesforce Commerce Cloud

We have added the following new tables:

• PRODUCT_IMAGE_GROUP
• PRODUCT_IMAGE_GROUP_IMAGE
• PRODUCT_IMAGE_GROUP_VARIATION_ATTRIBUTE
• PRODUCT_IMAGE_GROUP_VARIATION_ATTRIBUTE_VALUE

SAP Concur

We have added a new column, report_id, to the ITEMIZATION table.

Snapchat Ads

We've added two new columns, objective_v2_type and promotion_type, to the CAMPAIGN_HISTORY table.

Stripe

We have added two new tables, CHECKOUT_SESSION_CUSTOM_FIELD and CHECKOUT_SESSION_CUSTOM_FIELD_DROPDOWN_OPTION, to store data from the custom fields of the Checkout Session object. We are gradually rolling out this change to all existing connections.

Tableau

Walmart Marketplace

Improvements

Adobe Workfront

You can now specify the Adobe Workfront environment whose data you want to sync when setting up the connection. To support this enhancement, we have added a new field, Environment, to the connection setup form. For more information, see our setup instructions.

Amazon Selling Partner

We have changed the data type of the following columns in the ITEM_DIMENSION table from BIGINT to DOUBLE PRECISION:

• item_height_value
• item_length_value
• item_weight_value
• item_width_value
• package_height_value
• package_length_value
• package_weight_value
• package_width_value

We have made this change to preserve decimal values in dimension measurements instead of truncating them. We are gradually rolling out this change to all existing connections.

Apple App Store

We now support the Hybrid Deployment model for the Apple App Store connector. Learn more in the Apple App Store setup guide.

Azure Blob Storage

We have added a new authentication method, Microsoft Entra ID - Client credentials grant using client certificate. This authentication method enables you to securely connect using a Microsoft Entra ID service principal and a client certificate.

To support this feature, we have added the following fields to the connection setup form:

• Container name
• Storage account name
• Tenant ID
• Client ID
• Client certificate
• Private key for client certificate

For more information, see our setup instructions.

BigQuery destination

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the destination, which could cause differences between the source and destination data.

Databricks destination

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the destination, which could cause differences between the source and destination data.

dbt Cloud Orchestration

We've expanded the list of supported regions in the dbt Cloud orchestration setup form and added a new Service Provider selector.

Previously, only a limited set of regions were available. You can now choose from all current dbt Cloud regions, including cell-based regions that require an account prefix in the access URL.

To update your dbt Cloud orchestration connection to use a different region:

1. In the Fivetran dashboard, click Transformations.
2. Click Manage projects.
3. In the transformation list, click the three dots on the right side of the relevant dbt Cloud project.
4. Click Edit connection settings.
5. Select your Service Provider and Region.
6. Click Save & Test.

Email

We have added a new feature, support for files linked in the email body. You can now define a regex pattern to sync the attachments associated with your email. To support this feature, we have made the following changes to the connection setup form:

• Added a new option , File source, that allows you to choose between Attachment and URL in email body.
• Added a new field, URL link pattern, that allows you to sync files linked in the email body.

For more information, see our setup instructions

We are gradually rolling out this change to all existing connections.

Facebook Pages

You can now authenticate using a System User Token as an alternative to OAuth. This authentication method is recommended for organizations that don't want to ask their end-users to grant user access, such as digital marketing agencies or Powered by Fivetran users.

For more information about setting up this authentication method, including permissions and access level requirements, see our Facebook Pages setup guide.

Healthie

We have changed our sync strategy for the FORM_ANSWER_GROUP table. We now incrementally sync this table and use a 7-day lookback window to capture delayed updates.

JSON Double Quotes

Effective August 14, 2026, we have improved Fivetran replication for JSON scalar values for all connections. New connections created on or after August 14, 2026, will no longer persist these values with quotes (unless they are strings) in destinations, improving data consistency. Existing connections continue to use their current behavior.

Marketo

Our Marketo connector now supports an additional sync of activity data from the previous three days during each incremental sync, to help capture late-arriving activity records. To support this enhancement, we have made the following changes:

• Added a new toggle, Sync Late Arriving Activities, to the connection setup form
• Added a new parameter, trailingSyncEnabled, to the Fivetran REST API

For more information, see our Marketo documentation.

Microsoft Advertising

We updated how we process the view_through_revenue column. Microsoft Advertising returns these values with a comma as a thousands separator (for example, 1,013.99). We now remove the separator and store the values as plain numbers (for example, 1013.99) for consistency with the connector's other metric columns.

This change applies to data synced going forward. If you rely on the previous formatting, update any downstream logic accordingly.

MongoDB

We no longer support the oplog incremental update method. If your connection was using the oplog incremental update method, syncs will no longer work.

MySQL

If a DATE or DATETIME column contains a value that isn't a valid date or datetime (for example, a zero date like 0000-00-00), we now preserve the original value by promoting the column to STRING instead of syncing it as NULL.

If you have type locking enabled on a connection, this promotion doesn't happen — invalid values are synced as NULL instead, and the column keeps its original type.

We now preserve the zero-sentinel value 0000-00-00 00:00:00 in MySQL and MariaDB TIMESTAMP columns by changing the destination column type to STRING. Previously, we synced this value as NULL.

If you enable type locking for your connection, we keep the existing column type and continue to sync the zero-sentinel value as NULL.

This change affects connectors created on or after August 13, 2026.

MySQL destination

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the destination, which could cause differences between the source and destination data.

PagerDuty

We have changed the sync strategy for the following tables and their child tables. We now incrementally sync these tables instead of re-importing them during every sync:

• AUDIT_RECORD
• EVENT_ORCHESTRATION
• PAUSED_INCIDENT_REPORTING_ALERT
• SCHEDULE
• SLACK_DEDICATED_CHANNEL
• TEMPLATE

We have changed the re-import strategy for the following tables and their child tables. We now re-import them once a day to capture updates and deletes. Previously, we re-imported these tables during every sync.

• ADDON
• BUSINESS_SERVICE
• EXTENSION
• EXTENSION_SCHEMA
• INCIDENT_WORKFLOW
• LICENSE
• RULESET
• TAGS
• VENDOR

Pendo

We now support history mode for the following tables:

• ACCOUNT_APPLICATION_HISTORY
• ACCOUNT_HISTORY
• APPLICATION_HISTORY
• FEATURE_ELEMENT_PATH_RULE_HISTORY
• FEATURE_HISTORY
• GUIDE_HISTORY
• GUIDE_STEP_HISTORY
• PAGE_HISTORY
• PAGE_RULE_HISTORY
• STEP_HISTORY
• STEP_POLL_HISTORY
• STEP_THUMBNAIL_URL_HISTORY
• TRACK_TYPE_HISTORY
• TRACK_TYPE_RULE_HISTORY
• VISITOR_ACCOUNT_HISTORY
• VISITOR_APPLICATION_HISTORY
• VISITOR_HISTORY

Salesforce

We have upgraded our Salesforce and Salesforce Sandbox connectors from version 65.0 to 66.0 of the Salesforce REST API. For more information, see Salesforce documentation.

SAP Business ByDesign

We have fixed an issue where columns with address data in the BUSINESS_PARTNER_ADDRESS table were synced as NULL even when valid data for these columns was available in SAP Business ByDesign. By default, we will update the data for this table in a subsequent sync, which will count towards free MAR.

SFTP

We now support authentication using a keypair while configuring your connection using Hybrid Deployment. Previously, the Login with keypair? option was available only for connections configured using the SaaS Deployment model.

We are gradually rolling out this change to all existing connections.

Snowflake destination

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the destination, which could cause differences between the source and destination data.

SQL Server

Effective August 5, 2026, we have improved Fivetran Teleport Sync change detection for nullable columns in SQL Server connections. New connections created on or after August 5, 2026, distinguish NULL values from non-NULL values when calculating row hashes, improving data consistency in your destination. Existing connections continue to use their current behavior.

Effective August 7, 2026, we have improved change tracking (CT) accuracy for computed columns in SQL Server connections. New connections created on or after August 7, 2026, exclude non-persisted computed columns from CT syncs if they use nondeterministic expressions that return DATETIME or INTEGER values, such as GETDATE().

Previously, CT syncs evaluated these expressions when processing historical changes. As a result, the synced value could differ from the value at the time of the change, causing incorrect data in your destination. Existing connections continue to use their current behavior.

To resume syncing an excluded computed column, update it to use a deterministic expression.

Effective August 20, 2026, we've improved change tracking (CT) history mode syncs for SQL Server connections. New connections created on or after August 20, 2026, handle temporarily unavailable transaction metadata without omitting affected rows.

Previously, delays in the availability of SQL Server transaction metadata could cause some rows to be omitted during historical data processing. Existing connections continue to use their current behavior.

SQL Server destination

All new connections set up on or after August 5, 2026 preserve null byte characters in string values. Previously, we removed null byte characters before writing string values to the destination, which could cause differences between the source and destination data.

Stripe

We have expanded the DISCOUNT table to store discounts associated with subscription items. We are gradually rolling out this change to all existing connections.

Workday Adaptive Planning

We have added a new configuration option, Select all, to the Dimensions field in the connection setup form to allow you to select or deselect all available dimensions at once while configuring custom reports. For more information, see our setup instructions.