Rudderstack Release Notes

RudderStack now supports Pinterest Audience as a Reverse ETL destination. Connect a Reverse ETL source, an Audience Builder audience, or a Rudder Lookout audience, and RudderStack keeps a Pinterest customer list audience in step with it: members are added when they join and removed when they leave.

Why we built this integration

The Event Stream Pinterest Tag destination sends conversion events. To target the people in a warehouse audience on Pinterest, you need them in a customer list audience instead, and Pinterest accepts customer list uploads only through a signed-in Pinterest account.

Pinterest Audience connects with your Pinterest account through OAuth, creates the customer list and its audience for you or syncs into one you already have, and keeps it current in mirror mode.

Key features

• Create or reuse audiences: Create a new customer list audience from RudderStack, or sync into an existing one in your ad account.
• Email and mobile advertising IDs: Map an email address, a MAID, or both. Mapping both gives a better match rate.
• Hashing built in: RudderStack normalizes and SHA-256 hashes the values before sending them, unless you tell it your columns are already hashed.
• Three ways to sync: Use a Reverse ETL source, the Audience Builder, or the Rudder Lookout Activate flow.

Get started

1. In RudderStack, connect a Reverse ETL source to Pinterest Audience, or open an audience and add a sync to Pinterest Audience.
2. Add a Pinterest account and sign in to Pinterest as a user with access to your ad account.
3. Select the ad account, then create a new audience or select an existing customer list audience.
4. Map your email and MAID columns, and enable the sync.

Pinterest needs at least 100 matched users before an audience can be used for targeting, and it can take 48–72 hours to process an upload.

See the Pinterest Audience documentation for connection settings, hashing, and limitations.

Original source

Features

• jobsdb: pending events gauge for the gw jobsdb (#7447) (fc539c4)
• processor: report event_filter drops per reason and stop its succeeded rows (#7454) (5caff53)

Bug Fixes

• destination: authenticate marketo bulk uploads with bearer token (#7510) (3eaffad)
• processor: exclude dropped jobs from tracked users and activation records (#7485) (526602c)
• router: enable live events payload masking by default (#7500) (8b5c22d)
• router: mask multi-request delivery payloads (#7476) (f65572e)
• router: mask reporting sample event secrets (#7497) (833d694)
• router: report destination status for v1 proxy deliveries in transformer_outgoing_request_count (#7482) (bba1c21)

Miscellaneous

• bump sqlconnect-go to v1.34.0 (#7479) (71e8cfa)
• stop setting in_reported_by on report rows (#7458) (814d149)
• sync release v1.90.0 to main branch (#7475) (d05fcbb)
• tooling: move rudder-server test MinIO to Silo (#7469) (935bf59)

Original source

RudderStack now supports the HubSpot Audience destination — you can use it to keep a HubSpot static contact list in sync with a warehouse table, SQL model, or Lookout audience. HubSpot Audience adds contacts who enter the source and removes contacts who leave it. This sync changes list membership only — it does not create or update HubSpot contacts.

HubSpot Audience is in Private Beta. Reach out to Customer Success to enable it for your workspace.

Why we built this integration

The Event Stream HubSpot destination creates and updates contacts and sends event data. HubSpot campaigns run from static contact lists, and those lists go stale when membership lives in the warehouse or in Lookout.

HubSpot Audience keeps one static contact list aligned with that membership. You attach an existing list or create an empty one during setup. Mirror mode adds people who join and removes people who leave, so the list stays current for campaigns.

Key features

• List membership only: The sync adds and removes contacts on a static list. It does not create or update contact records.
• Existing or new list: Select a static contact list that already exists, or create an empty one during setup. Active lists do not appear in the picker.
• Record ID, then email: A mapped HubSpot Record ID is used first. Email is used only when the Record ID is empty. A Record ID that HubSpot does not recognize does not fall back to email.
• Warehouse, Audience Builder, and Lookout: Connect a Reverse ETL source, sync an Audience Builder audience, or activate a Lookout audience.

Get started

1. Create a HubSpot Service Key with crm.lists.read, crm.lists.write, crm.objects.contacts.read, and crm.objects.contacts.write.
2. Connect a Reverse ETL source to HubSpot Audience, add an Audience Builder sync, or activate a Lookout audience.
3. Select an existing static contact list, or create a new one.
4. Map at least one identifier: HubSpot Record ID, email, or both.

See the HubSpot Audience documentation for scopes, list behavior, and what a sync does.

Known limitations

• A full sync adds everyone currently in the source and removes no one. Contacts who left before that run can stay on the list.
• An email that does not match a HubSpot contact fails on both add and delete. This sync never creates the contact.

Resources

• HubSpot Audience: Scopes, sync failures, and what happens when you stop a sync
• How membership is written to HubSpot: Insert, delete, and full sync behavior
• Connect Reverse ETL Source to HubSpot Audience: Connect a warehouse table or SQL model
• Sync Audiences to HubSpot Audience: Sync an Audience Builder audience
• How to Activate an Audience to a Destination: Activate a Lookout audience

Original source

Private Beta

Load event data into a Microsoft Fabric Warehouse. Each sync stages files in a Lakehouse on OneLake, then loads them into the Warehouse.

2 minute read

Date: Oct 16, 2026

RudderStack now supports Microsoft Fabric as a warehouse destination. You can load event data from your sources into a Fabric Warehouse.

The Microsoft Fabric destination is in Private Beta. Reach out to Customer Success if you’re interested in enabling this destination for your workspace.

Why we built this integration

Microsoft Fabric is an analytics platform for data engineering, data warehousing, and business intelligence. This destination loads your event data into a Fabric Warehouse so you can query it with the rest of your Fabric data.

Each sync stages files in a Lakehouse on OneLake, then loads them into the Warehouse. The Lakehouse and the Warehouse need to be in the same region. Tables follow the Warehouse Schema, so the column layout matches your other warehouse destinations.

Key features

• Service principal authentication: You connect with a Microsoft Entra service principal. That is the only supported account type.
• Schema updates: The namespace schema is created when it doesn’t exist. Tables are created and altered as the event schema evolves. You can’t change the namespace after you create the destination.
• Append or merge: Warehouse append is on by default. Turn it off to merge rows and avoid duplicates. Merging can increase warehouse operation time.
• Scheduled syncs: The default frequency is every 3 hours. You can set a UTC start time and an exclude window. The 5-, 10-, and 15-minute options are available only on certain plans.
• Cloud mode: Every source type connects in cloud mode. You can also configure consent management providers.
• Case-sensitive names: Fabric’s default collation keeps the original case of schemas, tables, and columns. Use the same capitalization when you query them.

Get started

• Create a Fabric Warehouse and a staging Lakehouse in the same region.
• Register a Microsoft Entra app and record the Tenant ID, Client ID, and client secret Value.
• Ask your Fabric administrator to allow service principals to use Fabric APIs, and assign the app the Contributor role or higher on each workspace that contains the Warehouse or Lakehouse.
• In RudderStack, add Microsoft Fabric as a destination, connect it to your source, and select the workspace and Lakehouse.

See the Microsoft Fabric documentation for the full access setup and connection settings.

Original source

Private Beta

Sync warehouse and Audience Builder audiences to Gainsight PX as a Boolean custom user attribute.

Available Plans

• Free
• Growth
• Enterprise

2 minute read

Date: Oct 14, 2026

We’re pleased to announce support for the Gainsight PX Audience destination. With this integration, you can now sync audience membership to Gainsight PX Audience. Connect a Reverse ETL source or an Audience Builder audience, and membership is written as a Boolean custom user attribute on users who already exist in Gainsight PX.

Gainsight PX Audience is in Private Beta.

Reach out to Customer Success to enable it for your workspace.

Why we built this integration

The Event Stream Gainsight PX destination sends events such as identify and track. Gainsight PX engagements and segments target user attributes, so a warehouse audience still needs a membership flag on each user.

Gainsight PX Audience writes that membership to a Boolean custom user attribute you create in Gainsight PX. The attribute is true while a user is in the audience and false after they leave. You can target members, and users who are not in the audience, from the same attribute. Each sync runs in mirror mode, so the flag stays current.

Get started

Create the Boolean attribute in Gainsight PX before you connect the destination. This destination doesn’t create users or attributes. Members who are not yet Gainsight PX users fail until they exist there.

• In Gainsight PX, go to Administration > Attributes > User and create an enabled Boolean custom user attribute. Note its API name.
• Go to Administration > REST API and create a key with Read and Write permission. Select the data center that matches your Gainsight PX URL: US for app, EU for app-eu, or US2 for app-us2.
• In RudderStack, connect a Reverse ETL source to Gainsight PX Audience, or open an audience and add a sync to Gainsight PX Audience.
• Enter the attribute API name and map one column to the Gainsight PX user ID, identifyId.

See the Gainsight PX Audience documentation for connection settings, validation checks, and limitations.

Questions? Let's figure it out together.

Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Original source

Klaviyo destination now supports API version 2026-04-15.

New connections use API version 2026-04-15 by default, while existing connections keep their saved API version until you change it via the API Version to use destination setting. API version 2024-10-15 remains selectable but is deprecated.

Important considerations

For API version 2026-04-15, the selected Consent channel must match an identifier in subscription events: Email consent requires an email identifier and SMS consent requires a phone identifier. The event fails if no selected channel can be resolved.

Resources

See the Klaviyo Setup Guide for more information on selecting an API version and configuring consent channels.

Original source

Features

• processor: add source_out reporting metric (#7435) (909fba9)
• processor: allow disabling event schemas v2 per source (#7429) (a8c12b2)
• processor: report dest_transformer diff rows for multiplexing destination transformations (#7446) (cd95c32)
• processor: report dest_transformer succeeded for destinations without a destination transformation (#7445) (21f7e41)
• processor: report user_transformer succeeded for destinations without a user transformation (#7440) (f424257)
• router: mask live events destination credentials (#7460) (5fa5fe2)
• warehouse: add Microsoft Fabric destination (#7441) (74940c2)

Bug Fixes

• batchrouter: do not abort warehouse jobs on a zero failure timestamp (#7426) (506f6d5)
• batchrouter: recover snowpipe streaming channels for tables dropped externally (#7434) (93cd81d)
• destination: encode Lytics bulk upload URL values (#7433) (08e6937)
• destination: retry bqstream transport failures (#7464) (cd3dc3e)
• warehouse: cap deltalake validation schema describes (#7444) (d119779)
• warehouse: keep snowflake copy credentials out of logs (#7442) (f91747b)
• warehouse: set explicit lineSep for deltalake csv copy (#7455) (cc22d18)

Miscellaneous

• migrate self-hosted arm64 docker build runners to gha-runner-scale-set (#7425) (5cdffde)
• processor: retire the destination-filter reorder, dedup and gw_ingested reporting flags (#7450) (1040511)
• pull openapi-generator through the ECR pull-through cache (#7449) (6e9f920)

Original source

Sync SQL audiences you build in Lookout from ClickHouse to your destinations. The source runs in mirror mode and is in Private Beta.

Available Plans

Free
Growth
Enterprise

2 minute read
Date: Oct 7, 2026

You can now sync SQL audiences from ClickHouse with the ClickHouse Reverse ETL Source. Connect ClickHouse in Rudder Lookout, build an audience, and activate it to a destination. Lookout automatically creates the source in your linked RudderStack workspace when you activate the audience.

The ClickHouse Reverse ETL Source is in Private Beta.
Reach out to Customer Success to enable it for your workspace.

Key features

• SQL audiences in mirror mode: Build the audience in Lookout and activate it to a destination.
• Read-only access to your tables: RudderStack reads the rows your audience selects and writes only to a separate _rudderstack database, where it keeps snapshots and sync state.
• Diffs computed in ClickHouse: Each sync snapshots the current audience on your server and compares it with the previous snapshot. A normal sync sends removed members, then changed members, then new members.
• ClickHouse Cloud or a single node: Connect ClickHouse Cloud, including a service with several replicas, or a self-hosted server on one node. The server must run version 26.3 or later and accept HTTPS on a public address.

Get started

A Lookout admin adds the warehouse connection.

1. Reach out to Customer Success to enable the ClickHouse source for your workspace. Then, follow How to Connect Your RudderStack Workspace to Lookout.
2. In Lookout, go to Settings > Warehouses, click Add account, and select ClickHouse.
3. Enter the connection settings. Check the Setup SQL under the form, and ask a ClickHouse administrator to run it if they haven’t already.
4. Click Validate and create.
5. Build an audience on the ClickHouse connection and activate it to a destination.

See the ClickHouse Reverse ETL Source for the grant script, connection settings, and what you can sync.

Resources

• Grant permissions: The user, the _rudderstack database, and the grants each sync needs
• How to Build an Audience with the Agent: Build the audience on your ClickHouse connection
• How to Activate an Audience to a Destination: Send the audience to a destination and start the sync

Questions? Let's figure it out together.
Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.
Slack community Join now
Email support Reach out

Original source

Bot Management is now generally available. Detect bot traffic on every plan, and flag or drop it on Growth and Enterprise.

Available Plans

• Free
• Growth
• Enterprise

Bot Management is now generally available. You can detect bot traffic on web sources before it enters your data pipelines, so downstream tools work from cleaner event data.

Bot detection is available in the Free, Growth, and Enterprise plans. You can flag and drop bot events in the Growth and Enterprise plans only.

Key features

• Detect bot traffic on every plan: Identify bot events on web sources from user agent patterns.
• Flag or drop bot events: On the Growth and Enterprise plans, forward bot events with a flag for downstream analysis, or drop them before they enter your pipelines.
• Workspace and source settings: Apply one bot management configuration across all web sources, or override it for a specific source.
• Bot event metrics: Review bot event volume from the Events tab of a source.

Get started

1. In the RudderStack dashboard, go to Govern > Bot Management.
2. Turn on Bot Detection for your workspace.
3. On a Growth or Enterprise plan, choose whether to flag or drop the detected bot events.

See the Bot Management documentation for configuration steps, supported web sources, and how detection works.

Resources

• How Bot Management works: Learn how detection, flagging, and dropping handle bot events.
• Configure Bot Management: Set workspace and source-level bot management.
• View bot event metrics: Monitor bot event volume for a source.

Questions? Let's figure it out together.

Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Slack community Join now

Email support Reach out

Original source

Custom Types are now generally available on the Free, Growth, and Enterprise plans. Build a reusable library of property rules and apply them across your Data Catalog.

Available Plans

Free

Growth

Enterprise

3 minute read

Date: Oct 6, 2026

Custom Types are now Generally Available on the Free, Growth, and Enterprise plans. Previously in Beta for Enterprise workspaces, custom types let you define a validation rule once in a central library and apply it to any number of properties in your Data Catalog.

Custom Types are Generally Available. Free workspaces can create up to 2 custom types, Growth workspaces up to 10, and Enterprise workspaces an unlimited number. See Plan limits.

Why Custom Types?

Advanced rules apply to one property at a time. If you manage hundreds of properties, you end up copying the same enum, date format, or object shape onto each one, then editing every copy when the rule changes. A custom type holds that rule in one place, so a single edit reaches every property that uses it.

Key features

• Reusable rules: Define a custom type once and select it as the data type of any property. The property inherits every rule on the type.
• Complex types: Build an array of another custom type, such as a list of allowed enum values, or an object custom type with its own required and optional sub-properties.
• Dashboard or code: Create custom types in the dashboard, or define them as YAML and deploy them with Rudder CLI. The Data Catalog API endpoints for custom types remain in Private Beta.
• Plan-based limits: Each workspace has a cap on the number of custom types it can create. See Plan limits.

Plan limits

Plan Custom types per workspace Free 2 Growth 10 Enterprise Unlimited

When your workspace reaches its limit, you can still edit existing custom types but cannot create new ones. See RudderStack pricing to compare plans.

If your workspace goes over its limit, for example after a downgrade to a plan with a lower cap:

• Existing custom types become read-only.
• You cannot create new custom types, or create or update properties that use a custom type.
• You can still delete custom types to get back under the limit.

If you manage custom types with Rudder CLI, rudder-cli apply returns the limit error from the server when your workspace is at or over its limit.

Get started

1. Go to Govern > Data Catalog, open the Properties tab, then click the Custom Data Types tab.
2. Click Add custom type, then set a name, data type, and rules.
3. Create or edit a property, set its data type to Custom data type, and select your custom type.
4. Add the property to a tracking plan to validate events against it.

See Custom Data Types for the full setup guide, including array and object examples.

Resources

• Custom Data Types: Create custom types and use them in properties
• Use Custom Data Types in Rudder CLI: Define custom types in YAML and deploy them as code
• Data Catalog API: Manage custom types programmatically (Private Beta)

Screenshots

Rules tab for a new string custom type

Questions? Let's figure it out together.

Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Slack community Join now

Email support Reach out

Original source

Beta

Describe the customers you want to reach in plain language and let an AI agent build the audience from your warehouse and activate it.

2 minute read
Date: Oct 6, 2026

Rudder Lookout is an AI-powered analytics and activation agent for marketing teams. Describe the customers you want to reach in plain language, inspect the audience it builds from your warehouse, and activate that audience to the tools where you run campaigns.

Rudder Lookout is currently in Public Beta.

Why Rudder Lookout?

Building an audience from warehouse data previously meant writing SQL, raising a ticket, or waiting on an analyst. Lookout lets marketers describe an audience directly, understand how each rule affects it, and activate it without handing the workflow off to another team.

Key features

• Plain-language audience building: Describe the people you want to reach, and the agent reads your warehouse, identifies the tables that hold those people, and drafts the audience.
• Rule-level inspection: Review the spread of a numeric column, the values a text column contains, and how many people each choice keeps; adjust a rule and see the count move with it.
• Activation through existing destinations: Send the audience to a destination already in your RudderStack workspace and keep it in sync from there.
• Durable artifacts: Revisit and share every saved result instead of losing useful analysis in a transient chat reply.
• Multiple ways to work: Ask questions in Chat, turn results into saved Dashboards, or let approved external AI agents use Lookout MCP.

Get started

1. Sign in to Lookout with RudderStack.
2. Choose the RudderStack workspace to link to a new or existing Lookout workspace.
3. Connect the warehouse that contains your customer data.
4. Open a chat, describe the audience you want, and review the rules and audience count.
5. Save the audience and activate it to a destination in your RudderStack workspace.

See the Rudder Lookout Quickstart for setup and activation details.

Resources

• Rudder Lookout Overview: How Lookout works, what it connects to, and what it produces
• Rudder Lookout Quickstart: Workspace, warehouse, audience, and activation setup
• Audiences and Activation: Build, refine, and activate warehouse audiences
• Rudder Lookout Key Concepts: Core concepts and terminology
• Integrations: Connections to RudderStack, warehouses, and external tools
• Security and Data Handling: Data access, credentials, and security controls
• Rudder Lookout FAQ: Answers to common Lookout questions

Questions? Let's figure it out together.
Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Slack community Join now
Email support Reach out

Original source

Ask RudderAI questions directly in the RudderStack dashboard with a page-aware side-panel chat assistant.

RudderAI in Dashboard is a side-panel chat assistant with the same core monitoring, debugging, data quality, and documentation capabilities as RudderAI in Slack. You can investigate the page or resource you’re viewing without switching tools.

RudderAI in dashboard is currently in Public Beta and is available for new signups only.

Why RudderAI in the dashboard?

Investigating a pipeline from the dashboard previously meant moving to Slack and restating which source, destination, or page you meant. The dashboard assistant carries the current page context into your question, so you can start with the resource in front of you and stay in the same workflow.

Key features

• Page-aware context: A context chip identifies the current dashboard page or resource, and you can toggle it on or off before asking a question.
• Multiple conversations: Create and switch between threads with their full history retained, and delete conversations you no longer need.
• Flexible viewing modes: Resize the docked side panel or expand it to full-screen mode for focused work.
• Rich response cards: Open clickable resource lists, review error details and sourced documentation answers, or select a follow-up action from the response.
• Background processing: Close the panel while a response is running without canceling it, then reopen the panel to see the completed answer.

RudderAI cannot create, modify, or delete workspace configuration. RudderAI can make mistakes. Verify responses before acting on them.

Get started

Organization admins control dashboard access with the Enable RudderAI toggle in Settings > Organization. This setting is on by default for self-serve organizations that have access.

1. Open the RudderStack dashboard.
2. Click Ask RudderAI in the sidebar footer.
3. Toggle the page context as needed, then ask a question in natural language.
4. Rate the response with 👍 or 👎.

See RudderAI in Dashboard for usage details and access controls.

Resources

• RudderAI in Dashboard: Dashboard access, interactions, and permissions
• RudderAI in Slack: Slack setup and conversation behavior
• RudderAI Key Capabilities: Supported tasks and example questions
• RudderAI Security and Compliance: Access controls, data handling, and workspace isolation
• RudderAI FAQ: Answers to common RudderAI questions

Questions? Let's figure it out together.

Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Original source

The OpenAI Ads destination is now generally available to all RudderStack workspaces.

Available Plans

Free

Growth

Enterprise

less than a minute

Date: Oct 5, 2026

OpenAI Ads is now Generally Available. You can send advertising conversion events in cloud mode and web device mode from any RudderStack workspace.

Get started

1. Open the RudderStack dashboard and add or select a source.
2. Select OpenAI Ads from the destinations list.
3. Select or create an OpenAI Ads account and enter the required credentials.
4. Configure your event mappings and connect the destination to your source.

See the OpenAI Ads Destination Setup Guide for more information on setting up the integration.

Questions? Let's figure it out together.

Join the RudderStack Slack community to connect with other users, customers, and the RudderStack team — or reach out for direct support.

Slack community Join now

Email support Reach out

Original source