November 2026
Connector SDK
Fivetran Connector SDK
We now support deferred truncate for Connector SDK connections. Refer to deferred truncate for more details.
fivetran-connector-sdk PyPI package
We have released version 3.1.0 of the fivetran-connector-sdk PyPI package. This release includes the following updates:
- Added support for deferred truncate, which marks a table's existing rows as deleted only after the sync completes successfully, ensuring that existing data is preserved if the sync fails.
- Added support for reading package defaults from
pyproject.tomlwhen resolving deployment values from multiple sources. For more information, see field resolution order. - Improved the fivetran run command with bug fixes and enhanced logging.
- Improved template search and overwrite prompts for the fivetran init command.
- Minor enhancements and improvements.
Managed Data Lake Service
Managed Data Lake Service
We have deprecated the Retain All Snapshots option for the Snapshot Retention Period setting in the destination setup form. You can no longer select it for new destinations, and you can't switch an existing destination to it in the setup form or with the REST API. If your destination already uses Retain All Snapshots, you don't need to do anything. Otherwise, select another retention period.
We are gradually rolling out this change to all existing Managed Data Lake destinations. For more information, see our snapshot retention period documentation.
Schema changes
commercetools
We have added a new column, purchase_order_number, to the ORDERS table.
Intercom
We have added the following new tables:
COMPANY_NOTE_HISTORYCONTACT_MERGECONVERSATION_EXTERNAL_REFERENCECONVERSATION_MONITOR_EVALUATIONCONVERSATION_SCORECARDDATA_CONNECTORDATA_CONNECTOR_EXECUTIONSIDE_CONVERSATION_PART
We have added the following new columns:
| Table | Columns |
|---|---|
ARTICLE_HISTORY | ai_chatbot_availability, ai_copilot_availability, ai_sales_agent_availability, body_markdown, created_by_id, exclude_from_article_suggestions, help_center_audience, scheduled_publish_at, scheduled_unpublish_at, and updated_by_id |
CONTACT_HISTORY | email_domain |
CONVERSATION_HISTORY | channel_current, channel_initial, sales_agent_outcome, sales_agent_participated, sales_agent_routing_outcome, source_email_message_metadata_history, source_email_message_metadata_message_id, and source_email_message_metadata_subject |
EMAIL_ADDRESS_HEADER | drop_reason |
TICKET_HISTORY | previous_ticket_state_id |
Lightspeed Retail X-Series
| Table/column | Change type | Old name | New name | Notes | |
|---|---|---|---|---|---|
REGISTER | |||||
| table | New Table | REGISTER | |||
Paddle
| Table/column | Change type | Old name | New name | Notes | |
|---|---|---|---|---|---|
ADJUSTMENT | |||||
| column | New Column | payout_total_retained_fee | |||
| column | New Column | total_retained_fee | |||
Productive
| Table/column | Change type | Old name | New name | Notes | |
|---|---|---|---|---|---|
EXPENSE | |||||
| column | Deprecated Column | service_id | |||
FOLDER | |||||
| table | New Table | FOLDER | |||
TASK_LIST | |||||
| column | Deprecated Column | board_id | |||
| column | New Column | folder_id | Data is available in the folder_id column of the TASK_LIST table. | ||
TIME_ENTRY_VERSION | |||||
| column | New Column | deal_id | Data is available in the deal_id column of the TIME_ENTRY_VERSION table. | ||
| column | New Column | person_id | Data is available in the person_id column of the TIME_ENTRY_VERSION table. | ||
SAP SuccessFactors
| Table/column | Change type | Old name | New name | Notes | |
|---|---|---|---|---|---|
JOB_APPLICATION_STATUS_AUDIT_TRAIL | |||||
| column | New Column | job_app_status_id | |||
| column | New Column | job_app_status_set_id | |||
| column | New Column | job_application_id | |||
Improvements
AI Connector Agent
The AI Connector Agent is now generally available.
Read our AI Connector Agent documentation.
Azure DevOps
We have made the following changes to improve incremental syncs for the COMMIT table:
- We now maintain a separate sync cursor for each repository, which reduces unnecessary re-fetching of historical commits.
- We now send the cursor to Azure DevOps in 24-hour UTC format with an explicit
Ztimezone designator. Previously, the cursor used a 12-hour format without a timezone, which could cause Fivetran to request commits up to 12 hours earlier or later than intended.
Oracle Fusion Cloud Applications
We have added a new toggle, Disable Effective Date Filtering, to the connection setup form to support the extraction of historical data without applying any filters for an effective date. You must set the toggle to ON if you want to disable Oracle BICC effective date filtering and include records outside Oracle's default effective date range across all datastores in the connection.
This configuration option is available only for new connections. Be sure to modify your configuration before creating a connection.
QuickBooks
We have fixed an issue that incorrectly deleted records in the TAX_RATE_DETAIL table when either the sales or purchase tax rate list associated with a tax code was empty. By default, we will re-sync the TAX_CODE and TAX_RATE_DETAIL tables to restore these records. This re-sync counts toward free Monthly Active Rows (MAR). We are gradually rolling out this enhancement to all existing connections.
Slack
For CHANNEL_MEMBER, we now sync members from all channels to the existing table. Previously, we synced only members from channels where the app is installed.
TikTok Organic
We have added a new field, TikTok API rate limit tier, to the connection setup form. The field lets you match the connector's request rate to the rate limit tier TikTok has approved for your developer app (Basic, Advanced, Premium, or Ultimate), so that your syncs can run faster if your app has a higher tier.
The field defaults to Basic, so existing connections continue to sync at the same rate. Select a higher tier only if TikTok has approved your app for it. If you select a tier that your app has not been granted, your syncs may fail with rate-limit errors.
We have also added a new parameter, api_rate_limit_tier, to the Fivetran REST API.
REST API Improvements
Modify Destination Schema Names Before the Initial Sync
The Update a Connection endpoint now accepts the destination_schema_names parameter. Previously, the endpoint ignored this parameter, so you could change the setting only in the dashboard.
You can set destination_schema_names to FIVETRAN_NAMING or SOURCE_NAMING only before the connection's initial sync. After the initial sync starts, you can only send the value that's already set. Attempting to change it returns an error.