Version 2026.46
3 breaking11 new20 changed2 retiring23 fixed
Facebook Ads: dma column removed¶
The temporary dma column is removed from Facebook Ads extracts. The comscore_market column, and comscore_market_code for Marketing Mix Modeling, hold the same values. Switch downstream systems that use dma to these columns. For more information, see Facebook Ads: Setup guide.
Google Ads lookups no longer accept mccs argument¶
The Google Ads customer and billing setup lookups no longer accept the single mccs argument, so API requests that use it stop working. Use mccs/type and mccs/choices[] instead. The OPTIONS response lists these arguments. For more information, see Google Ads: Setup guide.
Workspace permission required to read mapping tables¶
Reading mapping tables and their entries with API keys now requires the Workspace: Read-only permission on the key. Requests with keys that lack it are refused. Add the permission to integration keys that read mapping tables. For more information, see Configuring mapping tables in Management API.
Facebook Ads: campaign_promoted_object field¶
Facebook Ads Campaign-level data now includes a campaign_promoted_object field, which identifies Advantage+ Shopping (catalogue) campaigns. Campaigns without a promoted object show a blank value. For more information, see Facebook Ads: Setup guide.
Google Ads: Asset automation setting attribute¶
Google Ads Campaign resources have a new attribute, Asset automation setting. It reports each asset automation type, such as final URL expansion or text asset automation, with its opted-in or opted-out status. For more information, see Google Ads: Setup guide.
Google Display & Video 360: new metrics¶
Google Display & Video 360 YouTube reports can now include the View-through Store Visits metric. Reach reports can include Media Cost (Advertiser Currency), which requires the Advertiser Currency dimension. For more information, see Google Display & Video 360: Setup guide.
Pinterest: campaign_is_performance_plus field¶
The Pinterest Advertiser Delivery Metrics report (campaign, ad group and pin promotion levels) and the Ads Analytics report now offer a campaign_is_performance_plus field showing whether a campaign is Performance+. For more information, see Pinterest: Setup guide.
Quantcast: Currency column in Standard report¶
The Quantcast Standard report now includes a Currency column when currency-related metrics are selected, as the Extended report does. Accounts that reach the Quantcast rate limit are retried instead of skipped. For more information, see Quantcast: Setup guide.
Teads Ad Manager: Campaign IDs filter¶
Teads Ad Manager datastreams have an optional Campaign IDs setting. When set, the report is filtered to those campaigns, so metrics such as reach are calculated only across them.
TikTok Ads: file_name dimension in Creative report¶
The TikTok Ads Creative report now offers a file_name dimension that returns the file name of video and image creatives. For more information, see TikTok Ads: Setup guide.
Twitter Ads: Targeting criteria metadata type¶
Twitter Ads datastreams can now use a Targeting criteria metadata type. It fetches the targeting criteria of the line items in the selected accounts and campaigns. For more information, see Twitter Ads: Setup guide.
Edit button on schedule summary cards¶
Custom and Smart Schedule summary cards in the Fetch box now have an Edit button that opens the schedule dialog on that entry. Smart Bundle schedules no longer offer Edit and Delete without permission to change the datastream. For more information, see Manual, Smart, and Custom fetches.
Expand toggle in Custom Script edit dialogs¶
The Edit Python expression, Edit regular expression and Edit Source Code dialogs for Custom Script now have an expand/collapse toggle. When expanded, the dialog uses the full window width and most of its height. For more information, see Using custom script transformations.
Data Quality monitor issues in the API¶
API keys can now read monitor issues, issue details, analytics and extract or job group issues, and preview custom monitor rules. With the monitor write permission, they can acknowledge, reopen and mark issues as expected, and accept a schema into the baseline. For more information, see Managing Data Quality issues in Management API.
New endpoints:
GET
/api/v1/monitors/issues/POST
/api/v1/monitors/issues/acknowledge/GET
/api/v1/monitors/issues/analytics/daily/GET
/api/v1/monitors/issues/analytics/top-datastreams/GET
/api/v1/monitors/issues/analytics/totals/GET
/api/v1/monitors/issues/extracts/{extract_id}/GET
/api/v1/monitors/issues/job-groups/{job_group_id}/POST
/api/v1/monitors/issues/job-groups/{job_group_id}/acknowledge/POST
/api/v1/monitors/issues/job-groups/{job_group_id}/unacknowledge/POST
/api/v1/monitors/issues/unacknowledge/GET
/api/v1/monitors/issues/{issue_id}/POST
/api/v1/monitors/issues/{issue_id}/apply-schema/PUT
/api/v1/monitors/issues/{issue_id}/expected/POST
/api/v1/monitors/preview/
Adform (Buyer): fewer report requests¶
For Adform (Buyer) Buyer Stats datastreams with all campaigns and tracking campaigns selected, report requests are halved. This reduces daily quota errors. For more information, see Adform (Buyer): Setup guide.
HubSpot: deal stage names from Pipelines API v3¶
The HubSpot Deals report now gets deal stage names from the Pipelines API v3, because the older version shuts down on 4 December 2026. The dealstage column still shows stage names, so no action is needed. For more information, see HubSpot: Setup guide.
LinkedIn Organizations & Brands settings form¶
The LinkedIn Organizations & Brands settings form offers only the time interval, breakdown and field options valid for the selected report type, and picks valid defaults. Posts shows Lifetime as its only time interval. For more information, see LinkedIn Organizations & Brands: Setup guide.
LinkedIn Organizations & Brands: API version 202609¶
LinkedIn Organizations & Brands now calls LinkedIn API version 202609 instead of 202601, which is sunset in January 2027. No action is needed and existing datastreams keep their configuration. For more information, see LinkedIn Organizations & Brands: Setup guide.
Shopify (GraphQL): GraphQL Admin API version 2026-10¶
The Shopify GraphQL connector now uses version 2026-10 of the GraphQL Admin API. Report types and fields are unchanged, so no datastream setup needs to change. For more information, see Shopify: Setup guide.
Shopify: REST API version 2026-10¶
The Shopify connector now uses version 2026-10 of the Shopify REST API. Existing datastreams keep working, but output values may differ where the newer API version changed the returned data. For more information, see Shopify: Setup guide.
Wikipedia connector renamed¶
The Wikipedia connector is now named Wikipedia, previously Wikipedia Pageviews, matching the native Wikipedia connector. New setups use the new name; existing setups keep the old one.
Yahoo DSP settings form layout¶
The Yahoo DSP datastream settings form now uses the standardized layout, showing Seats, Dimensions, Metrics, Interval, Timezone and Currency consistently with other connectors. For more information, see Yahoo DSP: Setup guide.
Updated issues and activity URLs in API responses¶
A datastream’s issues URL and a workspace’s overview URL in API responses now point to the current issues and activity pages. The old URLs still redirect. For more information, see Listing datastreams in Management API.
Authorization step in datastream and destination wizards¶
The Choose an Authorization step of the datastream and destination wizards is now a new page, with the same choices and labels. Back routes from later steps return to this step, and the item is created in the workspace the wizard was opened in. For more information, see Creating a datastream.
Extract preview fits the page¶
The extract preview now fits the page, with the table scrolling inside it. The transformation preview opens in the same space with a Back to extract preview link. The View column statistics buttons are removed from the transformation preview.
File upload for file-based datastreams in a modal¶
Uploading a file to a file-based datastream (FTP, SFTP, cloud storage, OneDrive, SharePoint, Google Drive, Box, Amazon S3, Azure Blob Storage and Mailgun) now happens in a modal on the datastream page. Existing upload links and bookmarks keep working. For more information, see Manual, Smart, and Custom fetches.
Legacy page URLs redirect directly¶
The /home/ and /datastreams/list/ URLs now redirect to the same destinations as before. The retired /agents/<id>/ URL, which showed an empty page, now opens the workspace overview.
Monitor issue access limited to permitted workspaces¶
Monitor issues, analytics and acknowledgements now exclude workspaces the user cannot reach. Only roles that can view monitors may acknowledge issues or mark them expected; roles without monitor access can no longer do so through the API. For more information, see Viewing details of the Data Quality issues.
Old datastream and activity links redirect to new pages¶
Old datastream and activity page links now redirect straight to their new pages, and an activity link opens the workspace it names.
Rescue stall detection adjusted¶
Rescues that are still queued, reserved or finishing on a shutting-down worker are no longer failed as stalled. Rescues that run longer than twice the rescue stall timeout tenant setting (24 hours by default) are failed and their tasks stopped.
Stricter checks on database connection options¶
Unsafe SQL connection options, invalid hostnames and ODBC values are refused, so fetches or connections that use them fail. SQL Server ApplicationIntent and Connect Timeout options are now allowed. Invalid connection arguments are ignored with a job log warning. For more information, see Database: Setup guide.
Transformation import in a drawer¶
Importing transformations from an export file now happens in a drawer on the Transformations page, opened with the Import button. The old import URL redirects to the Transformations page. For more information, see Introduction to the Transformations page.
Universal monitor settings follow API rules¶
Universal monitor settings in the web app now follow the same rules as the API, for example a timeliness monitor is always a warning. The web app and the API behave the same for these settings. For more information, see Working with universal monitors.
Usage Manager unsubscribe page removed¶
The Usage Manager notification unsubscribe page is removed. Its old address redirects to the Usage Manager. No settings are lost, because the page never saved any.
TikTok Ads: GMV Max Campaigns report deprecated¶
The GMV Max Campaigns report type in TikTok Ads is deprecated. Existing datastreams keep fetching with a warning, but new ones cannot be created. Move to the TikTok GMV Max connector. For more information, see TikTok Ads: Setup guide.
Webgains: Transactions report type deprecated, new report added¶
The existing Webgains report type is renamed Transactions (Deprecated) because Webgains has deprecated its endpoint. Existing datastreams keep working. A new Transactions report type uses the current Webgains Platform API, requires a Merchant ID, and optionally takes Program IDs and product-level basket detail. For more information, see Webgains: Setup guide.
BigQuery duplicate rows with empty key values¶
Reloading BigQuery destinations that overwrite by key columns no longer duplicates rows when a non-text key column, such as DATE or INTEGER, contains empty values.
Brightcove: long refreshed access tokens¶
Brightcove authorizations can now save refreshed access tokens longer than the previous limit, so fetches blocked by the failed save work again.
Fetches hanging on unresponsive servers¶
Fetches and metadata updates no longer hang for up to 12 hours when a remote server stops responding, as with The Trade Desk and Amazon DSP. Outbound calls now time out and fail.
Google Ads Audiences: transfers failing after eight hours¶
Google Ads Audiences transfers now wait up to 24 hours for Google to process uploaded members instead of failing after about 8 hours.
Google Ads: Change Event fetch timeouts¶
Google Ads Change Event fetches no longer run until timeout when 10,000 or more changes share a timestamp. Further changes are skipped with a warning, and events are no longer duplicated across pages.
Google Search Ads 360: gzip failure with Record requests¶
Google Search Ads 360 and other streaming connectors no longer fail with a gzip error when Record requests is enabled. Fetches complete normally while recording.
LinkedIn Ads: default metrics for pivot-only datastreams¶
A LinkedIn Ads datastream pivoted only on CONVERSATION_NODE, CONVERSATION_NODE_OPTION_INDEX or OBJECTIVE_TYPE now gets default metrics and saves with a Daily, Monthly or Yearly breakdown.
Microsoft Ads: invalid column selections¶
The invalid DateType column is removed from Ad Performance reports, and Account and Campaign Performance column combinations that Microsoft refuses are rejected up front instead of failing the fetch.
Sklik: intermittent server errors¶
Sklik datastreams no longer fail intermittently with Too many items (HTTP 413) or other 5xx errors, because the connector now retries them.
Teads Ad Manager: trailing comma in Advertiser IDs¶
The Advertiser IDs setting for Teads Ad Manager now accepts a trailing comma and space instead of rejecting them.
YouTube Analytics: Key Date Column and daily splitting¶
New YouTube Analytics datastreams get the correct Key Date Column, so daily extract splitting no longer fails. Existing affected Video and Playlist datastreams are corrected, and Local Data Retention keeps the datastream’s own column.
Zendesk: failed fetches after temporary server errors¶
Zendesk fetches that resume after a temporary server error no longer fail. Refetch the range for fetches that failed earlier.
Column endpoints limited to accessible workspaces¶
Column, datastream settings, column overwrite and data schema endpoints now only read or change columns in workspaces the caller can access. Other columns return 404, and a bulk column PATCH naming one returns 400.
Duplicate target column mappings refused by the API¶
The API now refuses a datastream mapping that sends two source columns to the same target column. Such a mapping was previously accepted and then failed at load time.
Mapping tables created by transformations saved through the API¶
Creating or updating a Custom Script transformation through the API again creates missing mapping tables referenced by its map instructions. If a table cannot be created or accessed, the transformation is saved and the request returns a 400 error.
Schedule history on datastream creation through the API¶
Creating a datastream through the API with a schedules list now records one Schedule (added) History entry per schedule, as flat schedule fields already did.
Azure datalake pruning with missing credentials¶
Pruning no longer leaves files in an Azure datalake when its credentials are missing or rejected. Extracts are kept and deleted on a later run once credentials are fixed.
Custom schedule deletion dialog¶
Deleting a Custom schedule keeps the dialog open until the delete finishes, goes straight to the empty state after the last entry, and shows an error if the delete fails.
Duplicate volume issues when reprocessing extracts¶
Reprocessing a day-split extract no longer re-runs the volume check against the original fetch’s data, so duplicate volume issues and notifications no longer appear.
OAuth2 authorization forms timing out¶
The settings form and wizard Configuration step of OAuth2 authorizations no longer time out on tenants with many access tokens.
Scheduling switch without schedules¶
The Scheduling switch on a datastream with no schedules now shows as off and disabled until a schedule is added.
Streaming extract download errors¶
Downloading a streaming datastream extract no longer fails when there are no matching extracts or a file is empty; it returns an empty CSV or JSON file.
Workspaces scheduled for deletion in v1 API endpoints¶
The v1 API workspace list, usage, warehouse query and schema, extract requeue and bindings endpoints no longer reach workspaces scheduled for deletion. Their rows are also left out of warehouse query results.