October 2026

Connector SDK

Fivetran Connector SDK

We now support an optional setup form for Connector SDK connections. You can define form fields and setup tests in Python that Fivetran renders in the dashboard. This feature is in beta. Refer to our Connector SDK Setup Form documentation for more details.

fivetran-connector-sdk PyPI package

We have released version 2.13.0 of the fivetran-connector-sdk PyPI package. This release includes the following updates:

• Adds the configuration command to run the connector's configuration form with interactive prompting support and create configuration.json locally from the provided input. Supports an additional --test flag to run connector tests locally.
• fivetran debug now saves crash reports to the crash-reports directory of your project when your connector code fails.
• Improved CLI prompts and help text.
• Minor enhancements and improvements.
• Updated testers.

Lite Connectors

Lite connectors are now available for the following sources:

• Pennylane

Logs

Dynatrace External Logging - new log.source field

We now automatically tag logs with log.source: "Fivetran". This makes them visible in the built-in Dynatrace log source views and filters without additional configuration.

New Relic - Japan region support

You can now send Fivetran logs to a New Relic account hosted in the Japan region. To do so, in the New Relic setup form, select JP as the subdomain. Fivetran then sends logs to https://log-api.jp.nr-data.net/log/v1. No action is needed for existing New Relic log services.

Transformations for dbt

Automatic patch version updates for all dbt projects

Fivetran is gradually moving all dbt projects to automatic patch version updates. We run each moved project on the latest supported patch release of its dbt minor version, the same way we run projects with the Automatically use latest patch version toggle enabled. For example, a moved project that uses dbt 1.9.4 runs on the latest 1.9 patch release. Patch releases contain bug and security fixes for the same minor version.

For moved projects, the dbt project setup form doesn't show the Automatically use latest patch version toggle, and the dbt Core Version drop-down menu shows only minor versions.

You don't need to take any action. Your projects keep their current minor version. To move a project to a newer minor version, select it in the dbt Core Version drop-down menu.

dbt version support

We have added support for dbt version 1.12.5 for the following destinations:

• BigQuery
• Databricks
• PostgreSQL
• Redshift
• Snowflake
• SQL Server
• Azure Synapse
• Teradata Vantage

We have added support for dbt version 2.0.6 for the following destinations:

• BigQuery
• Databricks
• Redshift
• Snowflake

Minimum dbt version for new projects

New dbt projects now require dbt Core version 1.10.20 or later. This change addresses security advisories that affect earlier dbt Core versions. Existing projects can keep running their current version, but if you change an existing project's version, you must select 1.10.20 or later.

Schema changes

Apple Search Ads

We've added a new table, IMPRESSION_SHARE_REPORT. The table contains impression share data from Apple's Impression Share Reports.

Auth0

Chorus

Coupa

We have added a new column, reasons, to the APPROVAL table.

Facebook Ads

We've added a new table, PIXEL_EVENT_STATS. The table contains hourly Meta pixel event fire counts for the pixels used by your ad account.

The pixels are discovered from the CAMPAIGN_HISTORY, AD_SET_HISTORY, CUSTOM_AUDIENCE_HISTORY, and CUSTOM_CONVERSION_HISTORY tables. We sync PIXEL_EVENT_STATS only for the pixels found in rows that are fetched incrementally during a sync. To fetch historical pixel stats events, re-sync these tables.

Fivetran Platform Connector

We've added a new column, deleted_from_source_at, to the SOURCE_SCHEMA, SOURCE_TABLE, and SOURCE_COLUMN tables.

Forethought

Instagram Business

We have added two new columns, media_audio_type and is_ai_generated, to the MEDIA_HISTORY table.

We have added four new columns, story_link_clicks, total_like_count, total_comment_count, and total_view_count, to the MEDIA_INSIGHTS table.

Katana

Productboard

Profound

Reddit Ads

We've discontinued support for the gallery_item_caption column from all report tables, such as CAMPAIGN_REPORT, AD_REPORT, and CUSTOM_REPORT, because Reddit has deprecated this field.

SAP Concur

We have added a new column, address_code, to the VENDOR_GROUP and VENDOR_STATUS tables.

We have added the address_code column to the composite primary key of the following tables:

• VENDOR
• VENDOR_GROUP
• VENDOR_STATUS

We have added two new tables, EXPENSE_ATTENDEE_ASSOCIATION and EXPENSE_ATTENDEE_ASSOCIATION_CUSTOM_DATA.

Tableau

TikTok Ads

We've added smart_plus_creative_id to the SMART_PLUS_AD_CREATIVE_INFO table

We've added a new prebuilt report, CREATIVE_MATERIAL_REPORT_DAILY, which tracks per day performance metrics for individual creative materials used in Smart+ ad campaigns. The CREATIVE_MATERIAL_REPORT_DAILY table's primary keys are:

• advertiser_id
• smart_plus_ad_id
• main_material_id
• stat_time_day

The new prebuilt report offers standard creative performance metrics, including impressions, clicks, spend, video engagement, conversions, and social engagement (likes, comments, shares, follows, profile visits).

UserVoice

We've added the following new tables:

• INTERNAL_STATUS
• INTERNAL_STATUS_UPDATE_EXTERNAL_USER
• INTERNAL_STATUS_UPDATE_TEAM
• INTERNAL_STATUS_UPDATE_USERS

YouTube Analytics

We've added the following new tables:

• CHANNEL_REACH_BASIC_A1
• CHANNEL_REACH_COMBINED_A1
• CONTENT_OWNER_REACH_BASIC_A1
• CONTENT_OWNER_REACH_COMBINED_A1

Zendesk

We've added the user_segment_ids column to the ARTICLE table.

Improvements

Activations

On October 15, 2026, we are introducing a unified Fivetran Activations experience with the following improvements:

• Source-based navigation: Manage activations and syncs by source. Previously, activations were organized by individual syncs. The redesigned interface groups syncs by each unique source-to-destination path, similar to how we organize tables within Fivetran connections. You can more easily find individual activations, trial status, and settings.
• Unified access controls: Fivetran Users & Permissions replaces the separate Census access model. The Fivetran access model supports SCIM provisioning, team management, and more granular permissions. Learn more in our Role-Based Access Control documentation.
• Activations API V2: Use the new API version to support source-based management and unified access controls. Learn more in the API V2 documentation. We will retire Activations API V1 on January 15, 2027. If you still use API V1, you must migrate to API V2 before then.

As part of the move to Fivetran Users & Permissions, Workspaces are now deprecated. We will retire Workspaces on January 15, 2027.

Adobe Workfront

We've added Adobe Developer Console as a new authentication method for the Adobe Workfront connector. Adobe is retiring custom Workfront OAuth2 applications: you can no longer create new applications starting November 1, 2026, and existing applications stop working on February 1, 2027. We recommend that you switch your existing connections to the Adobe Developer Console authentication method before February 1, 2027. For more information, see the Adobe migration guide.

To support this enhancement, we have made the following changes:

• Added a new drop-down menu, Authentication type, to the connection setup form. The drop-down menu lets you select either Adobe Developer Console or the legacy Workfront OAuth2 application authentication method.
• Added the auth_type parameter to the Fivetran REST API. For more information, see the create a connection API reference.

For more information, see our setup instructions.

AI Reporting

We now offer a pre-built, dbt-compatible AI Reporting multi-source data model for AI usage and cost reporting. The model combines data from our Claude and OpenAI connectors.

Amazon S3

When using Dynamically extract tables, you can now click Auto-generate to have Fivetran use AI to suggest a table extraction pattern based on your file names. Review and adjust the generated pattern before saving. For more information, see our setup guide.

Asana

You can now connect to Asana using a Personal Access Token as an alternative to OAuth. For more information, see our Asana connector setup guide.

Azure Functions

We now validate the Function App name more strictly for Azure Functions connections. In line with Microsoft's Azure naming rules, the name must be 2-64 characters long, can contain only letters, digits, and hyphens, and can't start or end with a hyphen. If your Function App name includes Azure-generated, dot-separated parts (for example, <app-name>.<region>-01), those parts can contain only letters, digits, hyphens, and dots. This ensures Fivetran sends requests only to your Function App's azurewebsites.net address.

If your connection uses a name that doesn't follow these rules, the connection test or sync fails with an invalid input error. To fix it, enter only the Function App name exactly as it appears in the Azure portal, without https://, paths, ports, or query strings, in the Function App field of the connection setup form. For more information, see our setup instructions.

Claude

We now offer a pre-built, dbt-compatible data model for Claude. For more information, see our Claude data model documentation.

Concord

We have updated the sync strategy for the AGREEMENT table and its child tables. We now re-import these tables daily to capture updates and deletes. Previously, we incrementally synced them to capture new records.

Confluence

We now support Unstructured File Replication for our Confluence connector. To support this enhancement, we have added a new toggle, Sync attached files, to the connection setup form. For more information, see our setup instructions.

Databricks destination

You can now choose Saturday or Sunday for table maintenance on your Databricks destination on AWS or Azure. To support this enhancement, we have added a VACUUM Day drop-down menu to the destination setup form. The selected day applies to both VACUUM and OPTIMIZE operations. Existing destinations retain their maintenance day unless you change it.

The drop-down menu is visible when the Disable VACUUM operations toggle is set to OFF. For more information, see our Databricks table maintenance documentation.

Eloqua

We now format custom fields of type date in the ACCOUNT table using the ISO 8601 UTC format. For example, we now represent 2020-11-29 00:00:00.000 as 2026-09-06T02:00:19Z.

Previously, we formatted these values in Eastern time without a time zone indicator. We continue to store the values as strings.

Facebook Ads

Meta has deprecated the ids query parameter in the Graph API. Starting with Graph API v26.0, requests that use ids to fetch multiple objects at once return an error. As of October 27, 2026, this change starts applying to all remaining supported Graph API versions. For more information, see Meta's announcement.

The connector previously used the ids parameter to fetch some metadata entities in groups. We now request each entity individually, still grouped into batch requests. This change keeps your syncs working.

As a result, the connector sends more requests to the Facebook API during syncs. Syncs that fetch a large number of metadata entities may take longer. No action is required on your part.

Gainsight Customer Success

You can now authenticate your Gainsight Customer Success connection using OAuth 2.0, in addition to the existing API key authentication method. For more information, see our Gainsight Customer Success setup guide.

Instagram Business

We have upgraded our Instagram Business connector to version 26.0 of the Graph API. For more information, see the Graph API changelog.

Ironclad

We now support Unstructured File Replication for our Ironclad connector. To support this enhancement, we have added a new toggle, Sync attached files, to the connection setup form. For more information, see our setup instructions.

Lucca

We now sync records with dates up to two years in the future into the USER_LOCATION table.

Managed Data Lake Service

We’ve added a Storage region drop-down field to the Managed Data Lake Service setup form for all storage providers. When setting up a new destination, you can now select a region from the drop-down menu, which lists only regions supported by Fivetran. Choosing one of these regions helps us process your data closer to where it’s stored, which can help reduce data egress costs.

For existing destinations without a saved storage region, we have added an optional Storage region field that you can update.

We are gradually rolling out this change to all connections.

MNTN

We now offer a pre-built, dbt-compatible data model for MNTN. For more information, see our MNTN data model documentation.

OpenAI

We now offer a pre-built, dbt-compatible data model for OpenAI. For more information, see our OpenAI data model documentation.

Oracle Fusion Cloud Applications

We can now identify rows updated during a sync for all incremental tables in connections with Delete Capture Frequency set to Every day or Every week. To support this enhancement, we have added a new column, _fivetran_last_synced_ts, only to these incremental tables. By default, we have enabled the _fivetran_last_synced_ts column in your schema if you have configured your connection with the Allow all new data or Allow new columns schema setting. If your connection uses Block all new data, you may need to manually enable the column on the Schema tab. For more information, see our documentation. We are gradually rolling out this enhancement to all existing connections.

Our Oracle Fusion Cloud Applications connectors now support Oracle Cloud Infrastructure (OCI) Object Storage for storing BICC extract files. Previously, we supported storing the extract files only in Universal Content Manager (UCM). To support this enhancement, we have made the following changes to the setup form:

• Added the Output File Storage field, to choose between UCM (Universal Content Manager) or OCI Object Storage.
• Added the following fields to configure OCI Object Storage:
  • External Storage Name
  • OCI Region
  • OCI Namespace
  • OCI Bucket Name
  • OCI Access Key
  • OCI Secret Key

For more information, see our documentation. We are gradually rolling out this improvement to all existing connections.

Pinterest Ads

You can now create custom reports with hourly granularity on targeting breakdown levels, and select Product Item Targeting as a custom report breakdown level.

Pinterest supports only one targeting type for these reports, so the custom report setup form shows a new Targeting type field where you select the targeting type to sync. For more information, see our Pinterest Ads documentation.

PostgreSQL destination

We have consolidated the PostgreSQL destination setup into a single setup form for supported PostgreSQL implementations and PostgreSQL-compatible databases, including Neon and Supabase.

We are gradually migrating all existing PostgreSQL destinations to the new consolidated PostgreSQL destination.

Recurly

We have made the following updates to the COUPON_REDEMPTION_HISTORY table:

• The coupon_id column now captures bulk coupon IDs only. Previously, it captured unique coupon code IDs.
• The unique_coupon_code_id column now captures unique child coupon code IDs. Previously, these IDs were stored in coupon_id.
• We will re-sync the table and backfill it with data from July 1, 2026.

Row filtering support for child tables

You can now apply a row filter directly to a child table, not just its parent table.

Row filters on a parent table and its child tables are independent. A row filter on a parent table is not applied to its child tables, so rows filtered out of the parent table are not automatically filtered out of the child tables, and the reverse is also true. Define a separate filter on each table if you want filtering applied to both.

Adding, editing, or deleting a row filter on a child table also triggers a re-sync of its parent table so that both tables stay consistent. If the parent table is excluded from the sync, the child table is not re-synced and keeps its existing data.

For more information, see our row filtering documentation.

Salesforce

We have changed our sync strategy for the Salesforce tables with the Attribute or Attr suffix. We now re-import these tables.

Sigma Computing

We have added the skipPermissionCheck=true query parameter to the WORKBOOK table to ensure that all workbooks are synced, including those not owned by or shared with the credential owner.

Snowflake Native to Lakehouse Tables Migration

You can now use the Fivetran dashboard to migrate a connection from a Snowflake destination that uses Native tables to one that uses Lakehouse tables. Lakehouse tables store your data in a Managed Data Lake Service destination and allow you to query the tables from Snowflake.

• For more information about migrating connections, see How to Migrate a Connection from Snowflake Native Tables to Lakehouse Tables.