Test Unpublished Integration Drafts
The `PostToQueueWithTriggerIdRequest` schema now includes a `workflowId` field, allowing you to target specific, even unpublished, versions of your integrations for testing.
Google Cloud Radar
Real-time intelligence on unreleased Google Cloud API updates, breaking schema evolutions, and method additions caught directly from Discovery Service.
Email Alerts
Breaking changes instantly & Monday digests across monitored GCP services.
October 2026 Updates
Breaking Change & Enhanced Testing for Integrations
A breaking change has made the `eucHardeningDisposition` field read-only in `IntegrationVersion`. Additionally, new parameters and response fields are available for `TestIntegrationsRequest` and `TestIntegrationsResponse`, offering more control and insight into credential flows and EUC hardening during integration testing.
Discovery Engine: Link Feedback to Agent Tasks
Discovery Engine now allows linking conversation feedback to specific Sobi agent tasks via a new `taskName` field. Minor internal scope adjustments were also made for `completeQuery` methods.
Methods:
Breaking: Widget Config & ToolCall Ops
Discovery Engine v1alpha introduces breaking changes: the `autoModelId` field in widget configurations is now read-only, and the `name` parameter for the new `toolCalls.operations.get` method is strictly required. Additionally, a new `get` method for tool call operations and a `taskName` field for conversation feedback are now available.
Methods:
Discovery Engine: `autoModelId` Now Output-Only; Sobi Task Tracking Added
A breaking change makes the `autoModelId` property in `WidgetConfigUiSettingsModelConfigInfo` read-only. Additionally, `FeedbackConversationInfo` now includes `taskName` for tracking Sobi long-running agent tasks, and `completeQuery` methods see scope reordering.
Methods:
Dataplex removes Unstructured Data Profiling
Dataplex has removed all API surface related to unstructured data profiling and graph profiling. This includes schema properties for data discovery, data scans, and detailed graph profile configurations. Developers using these experimental features must update their integrations.
BigLake Views Gain IAM, 'resource' Parameter Now Required
BigLake now offers full IAM policy management for Views with new `getIamPolicy`, `setIamPolicy`, and `testIamPermissions` methods. Heads up: The `resource` parameter is now strictly required across the API, which is a breaking change.
Methods:
Looker Instances Expose Authentication Types
Looker instances now expose their configured authentication types, including Google Auth and Workforce Identity Federation, through a new `authType` field. This provides clearer insight into instance security configurations.
Cloud Storage: New Hard Delete Pause for Soft-Deleted Objects
Cloud Storage buckets now support a 'hard delete pause' configuration within their soft delete policy. This new feature allows administrators to temporarily prevent the permanent deletion of soft-deleted objects, offering an additional layer of data protection.
Methods:
Vertex AI: Custom Voices, Cache Control, and Breaking Schema Changes
Vertex AI v1beta1 introduces custom voice management, allowing creation and listing of personalized voices. It also adds granular cache configuration updates. Be aware of several breaking changes, including new read-only fields and a type change for `structuredPrompt` in video generation. A previous breaking change regarding `A2ATaskRecord` read-only fields has been rolled back.
Methods:
Vertex AI: New Serving Profiles, Breaking Changes & Scope Updates
Vertex AI introduces a new `ServingProfile` resource for managing model serving configurations. This update also includes several breaking changes, notably making `parent` and `name` parameters strictly required, and changing the type of `structuredPrompt` in vision models. OAuth scopes for `agents` methods have also been adjusted.
Methods:
Granular IAM for Dataplex Agent Tasks
Dataplex now exposes standard IAM policy management methods for `agentTasks` resources. Developers can programmatically get, set, and test IAM policies directly on their agent tasks, enhancing security and access control.
Methods:
AlloyDB Express Clusters & Breaking Observability Changes
AlloyDB v1beta now supports Express clusters via a new `expressConfig` property on the `Cluster` schema. Developers should note significant breaking changes: several observability and query insights properties on `Instance` and related schemas are now read-only.
AlloyDB Express Clusters & Observability Changes
AlloyDB v1alpha now supports creating Express clusters. Be aware of breaking changes: several observability and query insights fields on `Instance` are now read-only, becoming output-only information.
Breaking: AlloyDB Instance Observability & Query Insights Properties Are Now Read-Only
AlloyDB instances now expose `observabilityInstanceInfo` and `queryInsightsInfo` as output-only fields. Any existing code attempting to modify these properties will now encounter errors, as they are strictly immutable. Update your integrations to treat these as informational.
Methods:
YouTube Comments: `imageUrl` Now Output-Only
The `imageUrl` property in `CommentSnippet` is now read-only, meaning developers can no longer set it. Additionally, several media upload endpoints now accept a broader range of content types.
Methods:
Automate SQL Server DDL Handling
Datastream now provides new configuration options for SQL Server change tables, enabling automated management of capture instances when DDL changes occur. This simplifies handling schema evolution.
Gemini Notebook Audit Logs Now Available
The Google Workspace Admin Reports API now supports auditing for Gemini Notebook activities. Developers can use the `activities.list` method to retrieve these new reports.
Methods:
September 2026 Updates
Pub/Sub Bigtable Sink: Custom Row Keys & Columns
The Pub/Sub Bigtable integration now offers significantly enhanced control over data ingestion. Developers can define custom row keys and map message fields to specific Bigtable column families, moving beyond the default fixed schema.
Cross-Cloud Caching for External Connections
BigQuery connections to AWS, Azure, and Salesforce Data Cloud now support an optional cross-cloud caching feature. This can improve query performance and reduce network egress costs for external data sources.
Dataproc: Multi-Zone Clusters & Read-Only Field Updates
Dataproc now supports multi-zone cluster configurations, enhancing resilience and availability. This release also includes breaking changes, making several fields in `ApplicationAttemptInfo` and `VirtualClusterOperationMetadata` read-only.
Introducing Spend Caps for Billing Budgets
Cloud Billing Budgets now supports a new `spendCap` feature, allowing you to define and enforce spending limits directly within your budgets. This introduces a new `SpendCap` schema with `inputState` for control and `outputState` for monitoring.
New Object Context API & Breaking Changes Clarified
Cloud Storage introduces a new `objects.viewFullContext` method to retrieve detailed object context, defined by the new `ObjectFullContext` schema. While the diff reports breaking changes, these apply to the required parameters of this *new* method, not existing API calls.
Methods:
BigQuery: Secure Query Context, Native Parquet Geography Export, and a Breaking Change
BigQuery introduces a new `secureContext` for passing sensitive, context-specific information to queries, accessible via a new SQL function. Additionally, you can now export Parquet data with native Geography types. Heads up: `JobStatistics2.referencedLogicalViews` is now read-only, which may break existing integrations.
Breaking Change: AssistAnswer's `connectorDisplayNames` is now Read-Only
The `connectorDisplayNames` property in both `GoogleCloudDiscoveryengineV1alphaAssistAnswer` and `GoogleCloudDiscoveryengineV1betaAssistAnswer` schemas has been made read-only. This is a breaking change for applications that may have been writing to this field.
Discovery Engine: AssistAnswer Field Now Read-Only, New OAuth Auth Option
A breaking change makes `connectorDisplayNames` in `AssistAnswer` read-only, requiring updates if you were writing to it. Additionally, a new `basicAuthenticationEnabled` field provides more flexibility for OAuth integrations.
Discovery Engine: AssistAnswer `connectorDisplayNames` is now read-only
The `connectorDisplayNames` property in the `AssistAnswer` schema (v1 and v1alpha) is now `readOnly`. This is a breaking change for developers who were previously writing to this field, which provides human-readable names for internal connector agents.
AlloyDB: `trackActiveQueryPlan` Deprecated
The `trackActiveQueryPlan` property within AlloyDB's `ObservabilityInstanceConfig` schema has been deprecated. Developers should plan to migrate away from using this feature.
AlloyDB: Observability Config Field Deprecated
The `trackActiveQueryPlan` field within the `ObservabilityInstanceConfig` schema has been marked as deprecated. Developers should update their configurations to avoid using this field.
Methods:
AlloyDB Enhances IAM Group Authentication
AlloyDB now supports new `userType` enum values for IAM group-based authentication, enabling more flexible and granular control over database access for groups, users, and service accounts.
Methods:
YouTube Data API: `postId` Removed, `favorite`/`like` Activities Gone, `subscription` Restored
The YouTube Data API v3 has undergone significant breaking changes. The `postId` parameter and properties have been removed from comments and comment threads. Additionally, `favorite` and `like` activity types and their associated schemas are no longer available. On a positive note, the `subscription` activity type is no longer deprecated.
Methods:
Drive API clarifies concurrent permission warnings
The Google Drive API's documentation for `permissions.create`, `delete`, and `update` methods has been updated. The warning about concurrent permission modifications is now more explicit, clarifying its scope.
Methods:
Permission Concurrency Warning Clarified
The Google Drive API's permission methods (`delete`, `insert`, `patch`, `update`) now feature updated descriptions that clarify the scope of concurrent modification warnings.
Methods:
Major Breaking Changes: Data Taxonomy & Attributes Removed
The Dataplex API has undergone significant breaking changes, with the complete removal of the Data Taxonomy and Data Attribute Binding feature sets. Developers using these features must update their integrations immediately. A new IAM role configuration for Data Product Access Groups has been introduced.
Methods:
Increased Deobfuscation File Upload Limit
The Android Publisher API now supports larger deobfuscation file uploads, with the maximum size increased from 1.6 GB to 2.0 GB. This update benefits developers with larger apps.
Methods:
Client Telemetry for Pub/Sub Publish Operations
Cloud Pub/Sub introduces new schemas, `PublishOperation` and `PubsubClientTelemetry`, to provide more detailed client-side telemetry for publish operations, including hedging attempts and client-side start times.
Pub/Sub Adds Client Telemetry for Publish Operations
The Pub/Sub API now includes new schemas for client-side telemetry, `PubsubClientTelemetry` and `PublishOperation`. These additions provide insights into publish operations, such as hedging attempts and client-side publish start times, to aid in debugging and performance analysis.
Pub/Sub Adds Client Telemetry for Publish Operations
New `PubsubClientTelemetry` and `PublishOperation` schemas provide client-side debugging information for publish calls, including hedging attempts and start times.
Lakehouse Proxy Config Introduced (Breaking)
Dataproc Metastore now exposes `LakehouseProxyConfig` for managing phased migrations to Lakehouse Iceberg REST Catalogs. This new configuration and its properties are read-only, which is a breaking change for any clients attempting to write to them.
Metastore v1alpha: Lakehouse Proxy Config & Read-Only Fields
The Metastore v1alpha API introduces Lakehouse proxy configuration for phased namespace migrations. The new 'lakehouseProxyConfig' field in the Service schema and its sub-properties are now explicitly read-only, which is a breaking change.
Lakehouse Proxy Routing & Read-Only Config
Dataproc Metastore introduces new Lakehouse proxy routing capabilities for phased migrations to Iceberg REST Catalogs. The new `LakehouseProxyConfig` and its properties are now officially read-only, which is a breaking change if you were previously attempting to configure them.
Discovery Engine v1beta: License Configuration & Schema Doc Updates
Adds BillingAccountLicenseConfig schema properties and refines parameter documentation for suggestions, searches, and user events.
Discovery Engine v1alpha: Knowledge Spaces Operations & Sharing Config
Introduces operations get method for collaborative project knowledge spaces, updates agent sharing configuration scopes, and adds widget and billing license properties.
Methods:
Discovery Engine: Widget Config Component Tag & License Config Updates
Updates Discovery Engine v1 schemas, introducing read-only tag properties to WidgetConfigCollectionComponent and adding BillingAccountLicenseConfig.
Dataform: File path description clarified
A minor documentation update clarifies the reference point for file paths within Dataform workspaces. No functional changes or developer action required.
New `updateId` for Hosted App Responses
The `UpdateAppStoreHostedAppResponse` now includes an `updateId` to uniquely identify and track specific versions of hosted app updates, enabling correlation with publishing signals.
Analytics Hub: Custom Bigtable Row Keys & Column Mappings
Analytics Hub now provides advanced customization for Pub/Sub to Bigtable ingestion. Developers can define custom row keys using delimited fields or structured schemas, and control how JSON message fields map to Bigtable column families, offering greater flexibility for data modeling.
Detailed Export Telemetry and Self-Hosted Source Support
Looker's export process now provides granular telemetry via new ExportMetrics and ComponentMetrics schemas, alongside support for self-hosted instance sources.
Dataplex adds LLM token tracking for Data Documentation
Dataplex now provides granular observability for LLM-powered data documentation scans, including detailed token usage and compute unit consumption.
Agent Patching goes Async and Video 'Pro Edit' arrives
Vertex AI v1beta1 shifts Agent updates to Long Running Operations and introduces experimental 'Pro Edit' video capabilities. New observability features and Reasoning Engine garbage collection also debut.
Methods:
Vertex AI adds Video 'Pro Edit' and Agent Observability
Vertex AI introduces experimental video editing capabilities and OpenTelemetry-based observability for Agents. Notably, updating an Agent via patch is now a long-running operation.
AlloyDB adds Postgres 19 support and Cluster Editions
AlloyDB v1beta introduces support for PostgreSQL 19 and a new Edition classification for clusters, including a Developer edition. Note that the edition field on Backups is now read-only.
AlloyDB adds Postgres 19 support and Cluster Editions
AlloyDB v1alpha introduces support for Postgres 19 and a new edition classification for clusters, while making the edition field read-only on backups.
Clarified Unmanaged Account Resolution Logic
The Directory API documentation now explicitly details the HTTP response codes returned when resolving unmanaged user account conflicts during user creation.
Granular IAM Filtering and Search-As-You-Type Tuning
Discovery Engine adds resource-level IAM filtering for Gemini Enterprise and dynamic search-as-you-type scoring overrides. Alpha users also gain support for file attachments in long-running agent tasks.
Gemini Access Controls and Search-As-You-Type Tuning
This update introduces resource-level access controls for Gemini Enterprise, per-request Search-As-You-Type overrides, and a breaking change to widget model configurations.
Search-as-you-type tuning and IAM-based resource filtering
Discovery Engine adds granular scoring controls for Search-as-you-type and resource-level IAM filtering for Gemini Enterprise. A breaking change makes the widget 'family' field read-only.
Methods:
Track Play Store status in AppStoreAppActiveApkSet
Developers can now explicitly flag if APK sets are already published on Google Play and provide a corresponding version code.
CMEK support for Query Templates
Analytics Hub now supports Customer-Managed Encryption Keys (CMEK) for Query Templates, allowing developers to encrypt template definitions using their own KMS keys.
Cloud SQL introduces Workload Capture and Replay
Cloud SQL adds a new suite of methods to capture production database traffic and replay it on test instances for performance and upgrade validation.
Strict schema rules and new methods for Workload Capture
The Cloud SQL Admin API introduces strict validation and immutability rules for Workload Capture schemas, alongside new methods to list, start, stop, and replay database workloads.
MigrationSummary endTime is now read-only
A breaking change in the Metastore v1beta API marks the MigrationSummary endTime field as read-only. The createTime field description has also been updated to clarify its role in the migration process.
MigrationSummary timestamps updated in Metastore v1alpha
The MigrationSummary schema now marks endTime as read-only and redefines createTime to reflect the start of metadata reads.
MigrationSummary schema updates and breaking changes
The MigrationSummary schema has been updated with a new read-only endTime field and revised semantics for createTime.
Dataproc deprecates diskType in favor of type for Hyperdisks
Dataproc's AttachedDiskConfig schema has deprecated the diskType field, introducing a new type field for specifying Hyperdisk configurations.
New extended data type URL for Object Custom Context
Cloud Storage has added an extendedDataTypeUrl field to the ObjectCustomContextPayload schema to better identify custom context data.
NotebookLM API Restored with Stricter Schema Constraints
The NotebookLM alpha API surface has been reinstated after a recent removal, now featuring stricter parameter requirements and immutable fields.
BigQuery adds GCS mounts for Python UDFs and native UUIDs
BigQuery now supports mounting GCS buckets as volumes for Python UDFs and introduces a native UUID data type for SQL schemas.
Minor terminology update in AgentTool documentation
A minor documentation update for Vertex AI Agents changes references from 'GCP' to 'Google Cloud' in the AgentTool schema.
Dataform adds BigQuery Unit Tests and End-User Auth
Dataform now supports native BigQuery unit testing and granular end-user authentication for workflow invocations. Note that several metadata fields for unit tests and auth are now strictly read-only.
Dataform adds BigQuery Unit Testing and End-User Auth
Dataform introduces native support for BigQuery unit tests and granular execution modes. Several fields related to test results and authentication are now strictly read-only.
Data Protection Policies and Session Management Controls
Vertex AI Agent Builder adds native Sensitive Data Protection integration for data stores and connectors, alongside new session management controls and model 'thinking' levels.
NotebookLM Alpha Removed & New Data Protection Policies
This update includes a massive breaking change removing the NotebookLM alpha API surface while introducing new Sensitive Data Protection integrations and session management controls.
Data Protection Policies and Session Management Controls
Discovery Engine introduces Sensitive Data Protection (DLP) integration, configurable session management for Engines, and a breaking change making icon URLs read-only.
Methods:
Deprecation of Public/Private App Visibility Permission
A specific developer account permission related to public/private app visibility has been deprecated and is no longer supported.
Clarification on AppStoreAppActiveApkSet versioning
The API documentation now explicitly states that all APKs within an AppStoreAppActiveApkSet must belong to the same app version.
Secret Manager integration for transfer parameters
Developers can now secure sensitive transfer credentials using Secret Manager instead of raw strings. This update also marks specific configuration fields as read-only, which may impact generated client libraries.
Vertex AI: Breaking Parameter Changes & Auth Scope Shift
Vertex AI v1beta1 enforces mandatory name and parent parameters and transitions many resource fields to read-only status. Additionally, the API is shifting its primary OAuth scope to the service-specific aiplatform scope.
Vertex AI shifts to granular OAuth scopes
Vertex AI is promoting the specific `aiplatform` OAuth scope as the primary authorization method across most resources, while finalizing the enforcement of the `name` parameter as strictly required.
Time-bound Role Assignments in Directory API
The Directory API now supports expiring role assignments, allowing administrators to set a specific timestamp for when a user's role should be automatically revoked.
Drive files.copy now supports copying comments
The Drive API now allows developers to include comments when duplicating files via the files.copy method.
Minor documentation update for Assignment principal
The description for the principal field in the Assignment resource has been updated for clarity and grammatical consistency.
Granular Scopes for Video Ratings
The videos.getRating method now supports the youtube.readonly scope, allowing for more restricted access to rating data.
Cloud Billing adds Agent Messaging and Task Management
The Cloud Billing API has introduced a new interaction model involving 'agents' and 'tasks', while enforcing stricter requirements on existing parameters.
Support for External Link Reporting and Game Reward Limits
The API now supports reporting for external content link transactions and introduces redemption limits for game-specific one-time product offers.
August 2026 Updates
Scoped Search and SQL Dialect Controls
Dataplex introduces scoped searching via a new contexts parameter and adds support for specifying SQL dialects in data documentation, alongside a breaking schema change.
Breaking Changes to Agents and Transcription Config
Vertex AI v1beta1 introduces several breaking changes, including mandatory resource parameters and the removal of the recently added transcription 'mode'. It also adds standard IAM policy management for Agents and new sandbox authorization methods.
Vertex AI Agents get IAM controls and lifecycle updates
Vertex AI introduces IAM support and enhanced deletion controls for Agents, alongside significant breaking changes to Sandbox networking and Audio Transcription configurations.
Deprecation of Container Combine and Move Tag Methods
Google Tag Manager has deprecated the `combine` and `move_tag_id` methods within the containers resource, signaling a shift in how container migrations are handled.
Methods:
Clarified field requirements for Drive Approvals
The Drive API documentation now explicitly notes that the approvals.list method requires the fields parameter to return detailed results.
YouTube Data API v3: Major Breaking Changes & Strict Validation
A significant update to the YouTube Data API v3 enforces strict requirements for core parameters and marks several video statistic fields as read-only. Developers must update their requests to include mandatory fields like part, id, and channelId.
Apps Script API enforces required IDs and expands scopes
The Apps Script API now strictly requires scriptId, deploymentId, and versionNumber in relevant methods, while significantly expanding its list of supported OAuth2 scopes.
Breaking: Library Agent enforces required parameters
The Library Agent API has updated its schema to enforce 'name' and 'parent' as strictly required parameters while adding standard GCP global parameters.
Gmail API: Strict Parameter Validation & CSE Immutability
The Gmail API has updated its discovery document to enforce strict requirement checks on core parameters and transition several Client-Side Encryption (CSE) fields to read-only status.
Drive API v3: Major Schema Hardening and Breaking Changes
Google Drive API v3 has introduced significant breaking changes by making numerous ID parameters strictly required and marking several schema properties as read-only.
Methods:
Drive v2 API: Major Validation Tightening & Breaking Changes
The Drive v2 API discovery document has been updated to enforce strict requirements on resource identifiers and mark several metadata fields as read-only.
Massive Schema Tightening and New Data Safety Method
Dozens of parameters are now strictly required and many schema fields have become read-only, alongside a new method for Safety Labels.
Breaking Changes and Workspace Studio Support
The Reports API has introduced breaking changes regarding required parameters and read-only fields, while also expanding support for Workspace Studio applications in audit logs.
Major Schema Hardening and Granular OAuth Scopes
The Directory API has undergone a significant hardening phase, making dozens of parameters strictly required and marking system-generated fields as read-only. It also introduces a wide array of granular OAuth2 scopes for more precise access control.
Breaking changes and parameter enforcement in Data Transfer
The Data Transfer API now strictly enforces required ID parameters and has updated its discovery surface to include standard Google API parameters and mTLS support.
CEL Conditions and Precedence for Reservation Assignments
BigQuery is introducing logic-based reservation assignments using Common Expression Language (CEL) and a new precedence system to resolve matching conflicts.
Smart Transcription, Multi-Speaker Config, and Agentic Video
Vertex AI adds a 'Smart' transcription mode for automated cleanup of audio transcripts and introduces multi-speaker configurations. Video generation is enhanced with a new 'Extend' task and agentic processing modes for dynamic frame navigation.
Vertex AI adds Response Management and Smart Transcription
Vertex AI introduces lifecycle management for endpoint responses and a new "Smart" mode for audio transcription that automatically cleans up disfluencies.
Git branch management arrives for Dataform Workspaces
Dataform Workspaces now support native Git branch operations including checkout, fetch, and delete, enabling better programmatic control over development environments.
Data Catalog adds support for Graph Nodes and Edges
Data Catalog now supports NODE and EDGE entry types, enabling metadata management for graph-based data structures.
Data Catalog adds NODE and EDGE entry types
Data Catalog now supports graph-based metadata with new NODE and EDGE entry types, allowing for more granular classification of graph-related assets.
Video Outpainting, Audio Control, and Monitoring Refinements
Vertex AI introduces advanced video generation capabilities including outpainting and audio control, while refining agent monitoring filters and deprecating legacy ASR fields.
Vertex AI adds Video Outpainting and Advanced Audio Control
Vertex AI expands its video generation experimental suite with new capabilities for outpainting, video-to-video transformations, and granular audio track management.
Looker adds ESA migration support to ExportMetadata
Looker's ExportMetadata schema now includes a field for specifying source BigQuery datasets during Elite System Activity (ESA) migrations.
New jobType for Materialized View Refresh assignments
BigQuery now supports assigning reservations specifically for automated materialized view refreshes, allowing for better resource isolation and prioritization.
Dataplex adds ChangeRequest comments and EntryLinkType events
Dataplex now supports reviewer comments on ChangeRequests and introduces a new event schema for tracking EntryLinkType lifecycle changes.
Vertex AI: Session Compaction and Transcription Refactor
Vertex AI introduces session compaction for long-running agent histories and standardizes speech transcription configurations, alongside a breaking change to traffic type enums.
Methods:
Vertex AI adds Session Compaction and Audio API updates
Reasoning Engines now support session history compaction to manage context windows via summarization. Audio transcription APIs are being streamlined with new language code fields, and evaluation runs gain CMEK support.
Support for Notebooks, SQL assets, and GCS snapshots
Dataform is expanding beyond standard projects to support single-file SQL and Notebook assets, alongside GCS-based repository snapshots for scheduled executions.
Dataform adds support for SQL and Notebook pipelines
Dataform is expanding beyond standard projects to support SQL and Notebook assets as first-class pipelines. New features also include GCS repository snapshotting for scheduled notebook execution.
July 2026 Updates
Deprecation of proposer field in QueryTemplate
The proposer field in the QueryTemplate schema is now formally deprecated in favor of primary_contact.
Vertex AI: Task Stores, Speech Transcription, and Model Proxies
Vertex AI expands agentic capabilities with Task Stores for decoupled task management and integrated speech-to-text transcription for generative models.
Methods:
Vertex AI adds Compact HTTP Forwarding and Serving Operations
Vertex AI introduces a 'compact' HTTP forwarding method for publisher endpoints and adds standard long-running operation management for Serving Profiles.
Methods:
TLS Configuration for BigQuery Connections
BigQuery connections now support granular TLS configuration, including support for Private PKI and custom CA certificates.
Explicit statement types and new data skew metrics
BigQuery now explicitly identifies the SQL statement type in query responses and provides new percentile-based metrics for detecting data skew in query stages.
Vertex AI: Memory Bank Expansion & Breaking Agent Changes
New event ingestion and profile retrieval methods for Memory Banks expand agentic capabilities, while breaking changes remove several fields from Agent and Tool result schemas.
Vertex AI adds Audio Transcription and Memory Event Ingestion
Vertex AI introduces native audio transcription schemas and expands agent memory capabilities with event ingestion. New controls for video generation rewriters and granular index endpoint updates are also now available.
New reservationGroupPath field for hierarchical tracking
The Reservation resource now includes a read-only reservationGroupPath array, providing the full lineage of a reservation within hierarchical groups.
Vertex AI: Breaking GenAI Schema Changes & New Retrieval Tools
This update introduces significant breaking changes by removing generation penalties and caching fields, while adding new proxy methods for model responses and expanding retrieval tool support for agents.
Methods:
Vertex AI adds HTTP request forwarding and video alignment
Vertex AI introduces a flexible HTTP forwarding method for publisher endpoints and adds experimental alignment controls for video generation and grounding observability.
Methods:
Breaking: context field removed from LookupContextRequest
The optional context string has been removed from the LookupContextRequest schema, affecting metadata lookup workflows.
Vertex AI adds Memory Bank CRUD and Streaming Event updates
Vertex AI expands its Reasoning Engine capabilities with full CRUD support for Memory Banks and introduces a distinction between legacy and step-based streaming events.
Vertex AI adds CRUD and Generation for Agent Memories
Vertex AI expands its memory management capabilities, allowing developers to create, list, and automatically generate persistent memories for AI agents within the Reasoning Engine framework.
Methods:
Table-valued arguments and multi-cloud storage stats
BigQuery now supports table-valued arguments for routines and provides granular object storage statistics for multi-cloud queries. Developers can also now force on-demand billing by setting reservations to 'none'.
Removal of metadataDestination from TransferRun
The output-only metadataDestination field has been removed from the TransferRun resource schema.
Vertex AI adds Claude-specific feature controls
Developers using Anthropic's Claude models on Vertex AI now have a dedicated configuration for advanced AI features and data retention opt-ins.
Vertex AI: New Memory Bank Resources and Agent Refinements
Vertex AI introduces MemoryBanks and Memories for persistent agent state management, alongside schema refinements for Monitored Agents and user scenario generation.
Vertex AI adds Memory management and Agent configuration updates
Vertex AI introduces long-running operation management for Memory resources and updates Agent scenario generation to support Gemini-based configurations.
Dataplex Data Domains add IAM policy member metadata
The recently introduced DataDomain resource now includes output-only IAM policy member strings, supporting both name-based and UID-based identity principals.
New Automatic Number Conversion and Error Schema Cleanup
Added automatic decimal separator detection for variable formatting and significantly reduced the surface area of the CompilerErrorLite schema.
Hierarchical Reservation Groups and Flex Plan Deprecation
BigQuery Reservation API adds support for updating reservation groups and hierarchical structures while deprecating Flex capacity commitments.
Methods:
June 2026 Updates
New score display modes for Lighthouse categories
The PageSpeed Insights API now includes metadata on how category scores should be visually represented, adding support for fractional scoring.
Vertex AI Beta: Computer Use Safety & Breaking Agent Changes
Vertex AI adds prompt injection detection for computer-use tasks and Zero Data Retention for Parallel AI Search, alongside significant removals of Agent-related methods.
Vertex AI introduces Agents and Semantic Governance resources
Vertex AI has significantly expanded its v1 API surface, introducing dedicated resources for managing AI Agents and a new Semantic Governance Policy Engine.
Methods:
Native Apache Arrow support for query results
BigQuery is introducing native Apache Arrow support for query results, enabling high-performance columnar data transfer directly from the query API.
Dataplex introduces Data Domains for logical organization
Dataplex adds a new DataDomain resource type, enabling developers to create, list, and manage logical data boundaries within a project and location.
Documentation cleanup and internal lint removal
Minor updates to parameter and field descriptions, primarily removing internal linting instructions and improving formatting.
Minor documentation cleanup and formatting
This update consists of minor documentation refinements, including the removal of internal linting comments and improved formatting for resource names in descriptions.
New Agentic Browsing category added to PageSpeed API
The PageSpeed Insights API now supports a new audit category called AGENTIC_BROWSING to evaluate site compatibility with AI agents.
New ingestion rule integrations for BigQuery and Airflow
The Data Lineage API now supports BigQuery and Managed Service for Apache Airflow (Cloud Composer) in ingestion rule configurations.
Python UDF concurrency documentation update
Updated documentation for Python UDF concurrency settings in ExternalRuntimeOptions, adding links to configuration guides.
Clarified transferredBytes for BigQuery Omni
The description for transferredBytes in query statistics has been updated to specifically define its role in BigQuery Omni data movement.
BigLake API rebranded to Lakehouse API
The BigLake API has been officially rebranded to the Lakehouse API in documentation, signaling a shift toward a unified metastore for multi-engine data lakes.
Deprecation of Destination methods in Tag Manager
Google Tag Manager has officially marked the 'get' and 'link' methods for the destinations resource as deprecated.
New Agentic Browsing Lighthouse Category
PageSpeed Insights has introduced a new 'agentic-browsing' category to its Lighthouse audit results.
Looker v1 adds Instance Undelete and Release Channels
Looker Core instances now support soft-deletion recovery and configurable update release channels (Rapid, Regular, Stable).
Dataplex adds Unstructured Data Profiling and Graph Profiles
Dataplex expands its scanning capabilities with support for unstructured data profiling and a new graph-based profiling schema for semantic inference.
Dataform introduces event-driven triggers and Git links
Dataform workflows can now be triggered automatically by table updates, and Git remote settings now support direct repository links for machine credentials.
Dataform adds GitRepositoryLink for machine credentials
Dataform now supports linking Git repository resources for machine credentials, simplifying authentication for automated workflows.
Generic HTTP request execution added to Connectors
Integration Connectors now supports a raw executeHttpRequest method, allowing developers to send arbitrary HTTP requests through existing connections.
Dataplex Universal Catalog integration for Data Transfers
BigQuery Data Transfer Service now supports automated metadata import into Dataplex Universal Catalog during data transfers.
BigQuery API clarifies IAM requirements across all methods
The BigQuery v2 discovery document has been updated to explicitly document the IAM permissions required for nearly every API method, simplifying least-privilege configuration.
Analytics Hub adds ZLIB compression for message transforms
Analytics Hub now supports ZLIB compression and decompression within message transformations, alongside updated row key logic for Bigtable subscriptions.
Vertex AI Beta: New Governance Engines and Breaking Agent Changes
This update introduces Semantic Governance Policy Engines and Agent Anomaly Detection while introducing several breaking changes to the Agents and Interactions APIs.
Online Evaluator Operations and GenerationConfig Deprecations
Vertex AI adds full operation management for Online Evaluators and marks several core GenerationConfig fields as deprecated in favor of a unified response format.
May 2026 Updates
Clarified location requirements for lineage search
The Data Lineage API now explicitly requires that the locations list in streaming search requests includes the location specified in the parent field.
Flexible Connector Parameters and Authentication
BigQuery connections now support arbitrary, connector-specific configuration and authentication parameters via a new map-based schema.
Methods:
Iceberg timestamp autodetection and UDF clarifications
BigQuery has expanded timestamp precision autodetection to Iceberg External Tables and clarified concurrency settings for Python UDFs.
Dataplex adds Change Requests for governed resource updates
Dataplex introduces a formal Change Request workflow, allowing developers to propose, list, and approve changes to data resources.
Methods:
Granular OAuth Scopes for Data Lineage Streaming
The Data Lineage API now supports specific read-only and read-write OAuth scopes for the searchLineageStreaming method, enabling better security posture.
Asynchronous repository deletion and Git branch improvements
Dataform introduces long-running repository deletion and clarifies Git default branch behavior.
Improved Git branch handling and path clarifications
Dataform now makes the Git default branch setting optional and provides a resolved 'effective' branch field, alongside clearer path definitions for directory entries.
