Google Cloud release notes

Stay organized with collections Save and categorize content based on your preferences.

You can now set a maximum slot limit for a reservation. You can configure the maximum reservation size when creating or updating a reservation. This feature is in public preview.

You can now specify which reservation a query uses at runtime, and set IAM policies directly on reservations. This provides more flexibility and fine-grained control over resource management. This feature is in public preview.

You can now allocate idle slots fairly across reservations within a single admin project. This ensures each reservation receives an approximately equal share of available capacity. This feature is in public preview.

Cloud Billing supports Dark theme in the Google Cloud console (in preview)

Dark theme is now available in the Billing section of the Google Cloud console (preview). To enable the Dark theme, in the Google Cloud console, click Settings > Preferences > Appearance. Choose Dark theme and click Save.

This release adds the Apigee Integrated Developer Portal Admin UI from the Classic Apigee UI into the Google Cloud console.

Leveraging Google Cloud console components provides API providers and Portal Admins with a centralized platform to efficiently configure, publish, and manage your API consumer portals, eliminating the need to switch between different UIs.

No new APIs have been introduced in this release.

See Publishing overview to get started.

You can now specify build dependencies in your build configuration file. For more information, see Manage build dependencies.

Cloud KMS with Autokey is now in General Availability (GA) for Cloud Run.

Public Preview: General purpose C4D machine types have reached Public Preview. C4D is powered by the fifth generation AMD EPYC processor (Turin) and Google Titanium.

C4D is designed to run mission critical workloads including web app and game servers, AI inference, web serving, video streaming, and data centric applications like analytics, relational, and in-memory databases.

C4D is available in standard, highmem, and highcpu machine types and only supports Google Cloud Hyperdisk storage.

To learn where to create C4D instances, see the Regions and zones page.

• Cortex for Meridian.
  • Cortex Framework for Meridian, Google's open-source Marketing Mix Modeling (MMM) tool (v1.0.5), delivers ready data models and automation of Meridian Model execution using Google Cloud tools like Enterprise Colab and Cloud Workflows.
  • This integration empowers users to make data-driven marketing decisions by providing accurate campaign performance measurement and budget optimization.
  • Cortex for Meridian simplifies the pre-modeling process by gathering and transforming data from core Cortex Framework data sources, including:
    • Cross Media & Product Connected Insights
    • SAP (ECC and S/4 HANA)
    • Oracle EBS

• Task Dependent DAGs:
  • Provided out-of-the-box, recommended DAGs and task dependencies for SAP ECC/S4 reporting.
  • Enabled the creation and deployment of customized task-dependent reporting settings for all data sources.

Dataproc Metastore multi-regional services now support the use of customer-managed encryption keys (CMEKs) -- (in preview).

Committed use discounts are now generally available (GA) for Firestore in exchange for a commitment to continuously spend a certain amount on Firestore read/write/delete operations for one year or three years. For details, see Committed use discounts.

Committed use discounts are now generally available (GA) for Firestore in Datastore mode in exchange for a commitment to continuously spend a certain amount on read/write/delete operations for one year or three years. For details, see Committed use discounts.

Looker (Google Cloud core) now supports Google group mirroring when using OAuth authentication.

Memorystore for Valkey supports storing and querying vector data. This feature is now Generally Available (GA). For more information, see About vector search.

Python 3.13 is now available in Preview.

Python 3.13 is now available in Preview.

Introduced logging and alerting capabilities to monitor the health and status of your backup/recovery appliances. You can configure email notifications via Cloud Logging to receive timely alerts on appliance status changes or potential issues.

Backup and DR Service now supports backup and restore of Db2 databases using persistent disk snapshots. This is typically faster and simpler than previous methods and in some cases may also reduce costs.

Introduced management console events for the Appliance Connectivity Events and Dynamic Protection Events.

You can now enable fine-grained access control on BigQuery metastore Iceberg tables. This feature is generally available (GA).

Support for the Python 3.13 runtime is now in Preview.

Cloud Run functions now supports the Python 3.13 runtime at the Preview release level.

The notebook gallery is now available.

The notebook gallery is a curated collection of notebooks to help you get started using Colab Enterprise. This collection consists of ready-to-use templates and examples to make it easier to learn new techniques, understand best practices, and get projects started quickly. Browse the notebooks by category or use the search bar to find a notebook that helps you get started. See the notebook gallery.

Generally available: Compute flexible committed use discounts (CUDs) are available for the sole-tenancy premium that you pay for eligible sole-tenant node types. Flexible CUDs add flexibility to your Compute Engine spending capabilities by eliminating the need to restrict your commitments to a single project, region, or machine series.

For more information, see Compute flexible CUDs.

Curated Detections has been enhanced with new detection content for Cloud Threats to include rule packs covering Office 365 and Okta. These rule packs are in public preview for customers with a Google Security Operations or Enterprise Plus license.

Curated Detections has been enhanced with new detection content for Cloud Threats to include rule packs covering Office 365 and Okta. These rule packs are in public preview for customers with a Google Security Operations or Enterprise Plus license.

Network Analyzer includes an insight that indicates if a GKE cluster's pod CIDR range isn't included in the ip-masq-agent ConfigMap. For more information, see GKE IP masquerade configuration insights.

The Execution: Ingress Nightmare Vulnerability Execution detector of Container Threat Detection is in Preview.

Chrome Enterprise Threats Category

This feature is currently in Preview.

Google SecOps has introduced a new detection category, Chrome Enterprise Threats, as part of the Curated Detections feature. This category provides rule sets for extension and browser threats. For more information, see Overview of Chrome Enterprise Threats Category.

Chrome Enterprise Threats Category

This feature is currently in Preview.

Google SecOps has introduced a new detection category, Chrome Enterprise Threats, as part of the Curated Detections feature. This category provides rule sets for extension and browser threats. For more information, see Overview of Chrome Enterprise Threats Category.

You can now manage backups for Memorystore for Valkey instances. This feature is Generally Available (GA).

Oracle Database@Google Cloud now lets you provision the Exadata Infrastructure instances with the new model X11M. This feature is generally available (GA). See Create Exadata Infrastructure instances.

The ability of Event Threat Detection to analyze foundational log sources is generally available (GA).

You can now use BigQuery DataFrames version 2.0, which makes security and performance improvements to the BigQuery DataFrames API, adds new features, and introduces breaking changes.

You can use partial ordering mode in BigQuery DataFrames to generate efficient queries. This feature is generally available (GA).

Airflow 2.10.5 is available in Cloud Composer.

In the Logs Explorer, you can now view the most frequently occurring fields and values in the JSON payload of your logs. For more information, see the Fields pane documentation.

You can now use custom constraints with Organization Policy to provide more granular control over specific fields for some Filestore resources. For more information, see Creating custom constraints for Filestore.

(New guide) Oracle E‑Business Suite with Oracle Database on Compute Engine VMs: Shows how to build the infrastructure to run Oracle E‑Business Suite applications with Oracle Database on Compute Engine VMs in Google Cloud.

GKE Inference Gateway is now available to significantly improve the performance, efficiency, and observability of generative AI workloads on GKE.

GKE Inference Gateway provides:

• Improved performance: AI serving tail latency is reduced, and AI serving throughput is increased through inference-optimized load balancing.
• Efficient resource utilization: Enables dense multi-workload serving of multiple LoRA fine-tuned models on a shared accelerator, leading to higher GPU/TPU utilization.
• Simplified operations: Features include model-aware routing, model-specific serving priority, and integrated AI Safety.
• Enhanced observability: Golden signals of observability are provided for inference requests.

Entity Context in Search

This feature enhances security investigations and incident response by letting users search for and view context events related to entities. It incorporates UDM entity context data to provide deeper insights into security incidents.

This feature is currently in Preview.

Entity Context in Search

This feature enhances security investigations and incident response by letting users search for and view context events related to entities. It incorporates UDM entity context data to provide deeper insights into security incidents.

This feature is currently in Preview.

Looker connector enhancements

You can now authorize the Looker connector to use the BigQuery OAuth credentials that you use with Looker, letting you view and interact with Looker Explores that use BigQuery data in Looker Studio. Learn more about how to authorize Looker data sources to use BigQuery OAuth credentials.

Custom organization policies are now generally available for Filestore. For more information, see Creating custom constraints for Filestore.

New troubleshooting tools for your service mesh are now available. You can get detailed error codes for your Istio resources and check the state of your mesh to identify and resolve configuration problems. Learn more about Resolving configuration issues and Understanding Feature State Conditions.

M129 release

• Updated the Dataproc JupyterLab plugin to version 0.1.85.

You can now create Memorystore for Valkey instances with the Cluster Mode Disabled configuration. This configuration is in addition to the Cluster Mode Enabled configuration that we support already. The Cluster Mode Disabled feature is available in Preview. For more information, see Enable and disable cluster mode.

M129 release

The M129 release of Vertex AI Workbench instances includes the following:

• Updated the Dataproc JupyterLab plugin to version 0.1.85.

Artifact Registry attachments are available in Preview for all repository formats. Attachments are artifacts that store metadata about a related artifact stored in Artifact Registry. To get started with attachments, see Store artifact metadata in attachments.

Regional endpoints are now available in Secure Source Manager. For more information, see Configure data locality by using regional endpoints.

Announcing data collectors data residency (DRZ) compliance for Apigee and Apigee hybrid.

Data collectors can be used with data residency for Subscription and Pay-as-you-go organizations and hybrid versions 1.14.0 and later.

See Data residency compatibility for information.

Announcing data collectors data residency (DRZ) compliance for Apigee and Apigee hybrid.

Data collectors can be used with data residency for Subscription and Pay-as-you-go organizations and hybrid versions 1.14.0 and later.

See Data residency compatibility for information.

Encrypting Dataplex data with customer-managed encryption keys (CMEK) is now available.

End-to-end tracing is now generally available (GA). Spanner now supports end-to-end tracing, along with client-side tracing in the Node.js and Python client libraries, in addition to Java and Go. For more information, see Trace collection overview.

Agent Assist offers a UI Connector with Salesforce to integrate with voice conversations.

(New guide) Harness CI/CD pipeline for RAG applications: Shows how to implement a continuous integration (CI) and continuous deployment (CD) pipeline for a retrieval-augmented generation (RAG) application in Google Cloud. The architecture uses CI/CD products from Harness to deploy containers to Cloud Run services.

The Apigee Extension Processor is now generally available (GA).

The Apigee Extension Processor lets Apigee customers add API management capabilities to Google Cloud and third-party products and services exposed using Cloud Load Balancing. Select from a range of Apigee policies that enable you to:

• Secure access to your workloads.
• Apply quota enforcement to network traffic.
• Manage Google access token and Google ID token injection to authenticate requests.
• Support native protocols like gRPC, SSE, and HTTP/3.

For more information, see the Apigee Extension Processor overview.

The Bigtable CQL client library for Java is available in Preview.

The Cassandra-Bigtable proxy adapter, which lets you connect your Apache Cassandra-based applications to Bigtable, is available in Preview.

The Bigtable Kafka sink, which lets you directly connect Apache Kafka and Google Cloud Managed Service for Apache Kafka, is now generally available (GA).

AlloyDB for PostgreSQL supports a 1 virtual central processing unit (vCPU) configuration with 8GB of memory, which is suitable for development and sandbox environments. For information about 1 vCPU supported regions and limitations, see Considerations when using 1 vCPU. This feature is generally available (GA).

AlloyDB supports AI-assisted troubleshooting that helps you resolve complex database performance issues like slow queries and high load. AI-assisted troubleshooting is available in Preview.

AlloyDB for PostgreSQL supports parameterized secure views, which provide a secure interface for application developers by improving data security and row access control while using SQL. This feature is in (Preview). For more information, see Parameterized secure views overview.

AlloyDB AI natural language (Preview) delivers secure and accurate responses for application end user natural language questions. For more information, see AlloyDB AI natural language overview.

AlloyDB AI query engine that builds on model endpoint management, and adds support for AI operators and Vertex AI multimodal and ranking models is available in (Preview). You can combine natural language phrases with SQL queries, like ai.if() for filters and joins, ai.rank() for ordering using ranking models, and ai.generate() for generating summaries of your data, and generate multimodal embeddings.

AlloyDB supports C4A Arm VMs on Google's custom-built Axiom processors. C4A VMs are available as predefined configurations from 1, 4, 8, 16, 32, 48, 64, and 72 vCPUs, up to 576 GB of DDR5 memory. C4A machines are available in limited regions. This feature is in Preview. For more information, see Considerations when using the 1 vCPU machine type.

AlloyDB now supports managed connection pooling in Preview. You can use managed connection pooling on your instances to improve the reliability, scalability, and performance of your workloads by optimizing resource utilization. For more information, see Configure managed connection pooling.

You can create and manage your App Hub applications using app-enabled folders, now available in Preview.

Gemini Cloud Assist in App Hub is supported in Preview. You can use the chat panel to retrieve information about your application in your app-enabled folder with Gemini assistance.

Gemini Cloud Assist for Artifact Registry is in Preview. You can learn about your container images with Gemini assistance.

To learn more, read the Gemini Cloud Assist overview.

You can now combine raster and vector data with the ST_REGIONSTATS geography function to perform geospatial analysis in BigQuery. For more information, see Work with raster data and try the tutorial that shows you how to use raster data to analyze global temperature by country. This feature is in preview.

You can now use the Apache Arrow format to stream data to BigQuery with the Storage Write API. This feature is available in preview.

Continuous materialized views for Bigtable are available in Preview.

SQL support for Bigtable is generally available (GA), including an UNPACK feature that lets you read time series data in a tabular format.

Logical views of Bigtable tables are available in Preview.

The Bigtable Studio query editor is generally available (GA).

Gemini-powered auto-conversion is now available in Preview for all heterogeneous migration scenarios. You can use code and schema conversion enhancements automatically provided by Gemini to significantly reduce the time and complexity of your database migrations.

For more information about auto-conversion and other AI conversion features, such as conversion assistant or pattern matching, see Accelerate code and schema conversion with Gemini.

Cross-Site Interconnect is available in Preview.

Cross-Site Interconnect is a new feature of Cloud Interconnect that helps you establish reliable, high-bandwidth Layer 2 connectivity between your on-premises network sites.

When you order Cross-Site Interconnect wires, Google provisions a transparent Layer 2 overlay over its global network between your two Cross-Site Interconnect locations.

For more information, see the Cross-Site Interconnect overview.

To help you get the right Cloud KMS keys on-demand, for consistent alignment with recommended encryption practices, Cloud KMS Autokey now has a free tier. The free tier covers the following usage:

• 100 free active key versions monthly
• 10,000 free cryptographic operations monthly

The free tier only applies to keys created using Cloud KMS Autokey. Key administration operations including key rotation are always free. For more details, see Cloud Key Management Service pricing

Application Monitoring lets you monitor the resources and infrastructure from the perspective of an App Hub application. The out-of-the-box (OOTB) dashboards generated for your application display log, metric, and incident data. These dashboards can help you understand how your application's resources are performing, and they can help you to diagnose issues. This feature is in Public Preview.

• Application Monitoring overview provides a brief overview of this feature.
• Set up application monitoring describes how to configure an observability scope so that you have an aggregated view of your log, metric, and trace data.
• View application telemetry describes the labels attached to your telemetry data, and it provides guidance about how to explore the OOTB dashboards.

Application Monitoring now supports app-enabled folders and App Hub host projects. For app-enabled folders, the metrics scope of the management project is automatically synchronized with the list of projects in the folder, provided quota is available. This feature is in Public Preview.

• Metrics scopes for app-enabled folders describes the synchronization algorithm and how to view your usage of the metrics scope quota.

Gemini Cloud Assist in Cloud Run is supported in Preview. You can use the chat panel to design, optimize, and troubleshoot your Cloud Run apps with Gemini assistance.

Cloud SQL Enterprise Plus edition now supports a new machine series called the C4A machine series, which provides optimized price-performance and delivers predictable high performance for high demand Cloud SQL workloads. C4A uses a new type of storage called Hyperdisk Balanced, and offers up to 72 vCPUs and up to 576 GB memory. The C4A machine series is available in Preview.

For more information about the C4A machine series and its availability, see Machine series overview.

Query insights for Cloud SQL Enterprise Plus edition is now generally available (GA) for your Cloud SQL Enterprise Plus edition for MySQL instances. Query insights for Cloud SQL Enterprise Plus edition offers fine-grained metrics such as wait events and granular query plan samples for faster root-cause analysis and intelligent index recommendations.

For more information, see Use query insights to improve query performance.

Cloud SQL for Enterprise Plus edition supports AI-assisted troubleshooting. With AI-assisted troubleshooting, you can resolve complex database performance issues like slow queries and high load for your instances in a guided manner. To use AI-assisted troubleshooting, you need Gemini Cloud Assist and query insights for Enterprise Plus edition. AI-assisted troubleshooting is available in Preview.

Cloud SQL Enterprise Plus edition now supports a new machine series called the C4A machine series, which provides optimized price-performance and delivers predictable high performance for high demand Cloud SQL workloads. C4A uses a new type of storage called Hyperdisk Balanced, and offers up to 72 vCPUs and up to 576 GB memory. The C4A machine series is available in Preview.

For more information about the C4A machine series and its availability, see Machine series overview.

Query insights for Cloud SQL Enterprise Plus edition is now generally available (GA) for your Cloud SQL Enterprise Plus edition for PostgreSQL instances. Query insights for Cloud SQL Enterprise Plus edition offers fine-grained metrics such as wait events and granular query plan samples for faster root-cause analysis and intelligent index recommendations.

For more information, see Use query insights to improve query performance.

Cloud SQL for Enterprise Plus edition supports AI-assisted troubleshooting. With AI-assisted troubleshooting, you can resolve complex database performance issues like slow queries and high load for your instances in a guided manner. To use AI-assisted troubleshooting, you need Gemini Cloud Assist and query insights for Enterprise Plus edition. AI-assisted troubleshooting is available in Preview.

Query insights for Cloud SQL Enterprise edition and Cloud SQL Enterprise Plus edition is now generally available (GA) for Cloud SQL for SQL Server. You can also now view the query details, query plans, and statistical query execution charts for your top queries.

For more information, see Use query insights to improve query performance.

The Datastream API now supports streaming data to BigLake managed tables. For more information, see Stream data to BigLake managed tables (BLMT).

You can now use account connectors to connect your Google Cloud account with individual accounts on supported non-Google Developer Tools providers. This feature is in Preview.

You can now use Query insights to view query performance metrics for your database. This feature is in Preview.

Firestore is now available on Database Center. You can track your Firestore resources in the fleet inventory section and the resource table in the Database Center. You can also use Database Center to monitor the following health issues for your Firestore resources:

• No automated backup policy
• No point-in-time recovery

For more information about Database Center, see Database Center overview. For more information about health issues supported for Firestore, see Supported health issues.

You can now use Query insights to view query performance metrics for your database. This feature is in Preview.

Gemini Code Assist tools are in Preview. You can use tools to access external services from your IDE. To learn more about tools, see the Gemini Code Assist Tools overview.

Agent Development Kit (ADK) is now available in Preview. For more information, see Agent Development Kit.

Vertex AI Agent Engine

The following features are now available for Vertex AI Agent Engine in Preview:

• Agent Development Kit integration
• Example Store
• LlamaIndex Query Pipeline integration

The following features are now generally available for Vertex AI Agent Engine:

• Agent monitoring

Gemini Live API is now available as a public preview offering and has been updated with the following features:

• Support for responses in 8 voices and 31 languages using Chirp 3
• Updated UI support in Vertex AI Studio
• Expanded conversation session window
• Ability to extend conversation sessions
• Support to share your current screen with Gemini during conversations
• Transcription support for audio in and audio out
• Support to change or update the system instructions mid-session

For more information, see Gemini 2.0 Flash Live API.

Agent Garden is now available in Preview. For more information, see Vertex AI Agent Builder overview or go directly to Agent Garden in the Cloud Console.

Gemini 2.5 Pro is now available as a public preview offering.

For more information, see Gemini 2.5 Pro.

Grounding: Grounding with Google Maps is now available as a Public Experimental feature. For more information, see Grounding with Google Maps.

Grounding: Web Grounding for Enterprise is now Generally available. For more information, see Web Grounding for Enterprise.

(New guide) Optimize AI and ML workloads with Cloud Storage FUSE: Learn how to optimize performance for AI and ML workloads on Google Kubernetes Engine (GKE) by using Cloud Storage FUSE.

In the Chart Config Editor, you can save a configuration as a template so that you can reuse it in other visualizations or share it as a starting point for other users.

The classification for the version, versions, and page_events API endpoints have been changed from "Admin" to "N/A" in System Activity queries. These endpoints no longer count toward Admin API endpoint quotas.

The Druid JDBC driver has been updated from 1.22.0 to 1.25.0.

The Athena JDBC driver has been updated from 2.0.35.1000 to 2.1.5.1000.

The Dremio JDBC driver has been updated from 4.5.0 to 25.2.0.

The Spark Databricks JDBC driver has been updated from 2.6.34 to 2.7.1.

The Exasol JDBC driver has been updated from 6.2.3 to 24.2.1.

The Denodo JDBC driver has been updated from 8.8.0 to 9.1.3.

The Trino JDBC driver has been updated from 402 to 468.

A new Labs feature, Fast Dev Mode Transition, improves the performance of Development Mode on your instance by loading LookML projects in read-only mode until a developer clicks the Create Developer Copy button for the project.

The New Database Connection Setup feature is now out of Labs and generally available. This feature updates the Add/Edit Connection page with a modernized UI, enhanced validation, connection testing capabilities, and a comprehensive configuration summary. If you want to revert to the legacy connections workflow, you can enable the Use Legacy Connections Page legacy toggle.

The Content Validator scoping feature is now generally available for customer-hosted Looker deployments (the feature is already available for Looker-hosted deployments). This feature lets developers scope the validation to specific LookML projects and a specific content folder (including its subfolders, if any). This can improve the performance of the Content Validator.

The New Database Connection Setup feature is now generally available. This feature updates the Add/Edit Connection page with a modernized UI, enhanced validation, connection testing capabilities, and a comprehensive configuration summary. If you want to revert to the legacy connections workflow, you can enable the Use Legacy Connections Page legacy toggle.

Recommendations for Memorystore for Redis are now available at Database Center. With this release, Database Center displays health issues about the manageability and performance of Memorystore for Redis. This feature is in Preview. For more information, see Database Center overview and Database health issues.

Recommendations for Memorystore for Redis Cluster are now available at Database Center. With this release, Database Center displays health issues about the manageability and performance of Memorystore for Redis. This feature is in Preview. For more information, see Database Center overview and Database health issues.

Gemini Cloud Assist for Flow Analyzer is in Preview. You can generate SQL queries for VPC Flow Logs with Gemini assistance.

Model Armor and GKE integration

Model Armor now enforces security policies uniformly on generative AI inference traffic using a traffic extension. This applies to all application load balancers, including Google Kubernetes Engine Inference Gateway. This feature is in Preview. For more information, see Integration with Google Kubernetes Engine.

The Google Kubernetes Engine (GKE) Gateway supports using extensions to add custom logic into the load balancing processing path. For more information, see GKE extensions. This feature is in Preview.

You can configure Model Armor with Service Extensions to protect AI workloads on supported Application Load Balancers. For more information, see Callouts to Google services. This feature is in Preview.

You can use Gemini assistance to help you use system insights to optimize and troubleshoot Spanner resources. For more information, see Optimize and troubleshoot with Gemini assistance.

Spanner offers Cassandra compatibility with API support and new migration tools allowing seamless lift-and-shift migrations of Cassandra applications. For more information, see Migrate from Cassandra to Spanner.

General availability support for the following integration:

• Database Center

Action required: You can access Kubernetes operator 1.4.0 high availability (HA) improvements for automatic setup, failover, and healing capabilities starting with AlloyDB Omni 15.7.1 and later. To access these features, see "Migrate to the latest version of AlloyDB Omni on Kubernetes" for AlloyDB Omni 15.7.1 and 16.3.0.

BigQuery ML now offers a built-in TimesFM univariate time series forecasting model that implements Google Research's open source TimesFM model. You can use BigQuery ML's built-in TimesFM model with the AI.FORECAST function to perform forecasting without having to create and train your own model. This lets you avoid the need for model management.

To try using a TimesFM model with the AI.FORECAST function, see Forecast a time series with a TimesFM univariate model.

This feature is in preview.

You can now create, view, modify, and delete Apache Iceberg resources in BigQuery metastore. This feature is generally available (GA).

You can now connect BigQuery metastore to Apache Flink. This feature is generally available (GA).

(New guide) Oracle E-Business Suite with Oracle Exadata in Google Cloud: Shows how to build the infrastructure to run Oracle E-Business Suite applications with Oracle Cloud Infrastructure Exadata in Google Cloud.

Custom organization policies are now generally available for Identity-Aware Proxy. For more information, see Use custom organization policies.

• You can ask Gemini Cloud Assist about active incidents, their impact, and updates
• You can view incidents related to your App Hub applications in Cloud Hub. See instructions for displaying incidents in Cloud Hub.

Vertex AI Search: Stream Google Cloud Storage buckets to data stores

In addition to one time and periodic imports from Cloud Storage, you can stream unstructured data from Cloud Storage into a data store. This lets you serve results from the bucket to your users in near real time.

Streaming must be set up at the bucket-level (not at the folder- or file-level), and the bucket may only contain unstructured data.

For general information about creating data stores, see Create a search data store.

BigQuery data preparation is generally available (GA). It offers AI-powered suggestions from Gemini for data cleansing, transformation, and enrichment. BigQuery supports visual data preparation pipelines and pipeline scheduling with Dataform.

You can now create remote models in BigQuery ML based on Llama and Mistral AI models in Vertex AI.

Use the ML.GENERATE_TEXT function with these remote models to perform generative natural language tasks for text stored in BigQuery tables. Try this feature with the Generate text by using the ML.GENERATE_TEXT function tutorial.

This feature is generally available (GA).

Smart-tuning is now supported for materialized views when they are in the same project as one of their base tables, or when they are in the project running the query. This feature is generally available (GA).

Direct VPC egress support for Cloud Run jobs is now generally available (GA).

You can now configure Identity-Aware Proxy (IAP) for Cloud Run to secure your services with a single click from all ingress paths (in Preview).

Configuring GPU in your Cloud Run service is now generally available (GA).

Cloud Run Threat Detection is available in Preview.

Cloud SQL now supports the Enterprise Plus recommender. Based on your application workloads and resource utilization, the recommender helps you optimize performance by identifying SQL Server instances that might see performance improvements when upgraded to Cloud SQL Enterprise Plus edition.

New Beta resources (direct reconciler)

• ApigeeEndpointAttachment

• ApigeeEnvgroupAttachment

• ApigeeInstanceAttachment

• ManagedKafkaTopic

• SecureSourceManagerInstance

• SecureSourceManagerRepository

New Fields

• GKEHubFeatureMembership

  • Added spec.configmanagement.configSync.stopSyncing in version1.129.

• SpannerInstance.

  • Added spec.defaultBackupScheduleType field.
  • Added spec.labels field

Premium parsers

Specific high-volume parsers are now categorized as premium. Google aims to address customer issues related to premium parsers as quickly as possible, typically within a few days.

For a complete list of different types of parsers and the level of support that Google provides for each, see Manage prebuilt and custom parsers.

For a complete list of premium parsers, see Default parser configuration and ingestion.

Premium parsers

Specific high-volume parsers are now categorized as premium. Google aims to address customer issues related to premium parsers as quickly as possible, typically within a few days.

For a complete list of different types of parsers and the level of support that Google provides for each, see Manage prebuilt and custom parsers.

For a complete list of premium parsers, see Default parser configuration and ingestion.

Preview: You can now enable IAP directly on your Cloud Run services without configuring load balancers.

For more information, see Configure Identity-Aware Proxy for Cloud Run.

IPv6 subnet exchange is generally available.

You can use export filters to configure a VPC spoke to exchange IPv6 subnet ranges or both IPv4 and IPv6 subnet ranges. For more information, see VPC connectivity with export filters.

For Autonomous Databases, you now have the options to set up public network access and private network access in Google Cloud. This feature is generally available (GA).

Cloud Run Threat Detection is available in Preview.

Create a quick action (Preview)

Administrators can now predefine quick actions for analysts to execute directly within cases and alerts.

The Quick Actions widget can be added to default case and alert views, and customized alert views within playbooks.

For more information, see Create a quick action.

What's New in Google SecOps

At the top of your Google SecOps screen, click the question mark and select What's New to display the top five new features in the Google SecOps platform.

Create a quick action (Preview)

Administrators can now predefine quick actions for analysts to execute directly within cases and alerts.

The Quick Actions widget can be added to default case and alert views, and customized alert views within playbooks.

For more information, see Create a quick action.

Access Approval supports Document AI in the GA stage.

Access Approval supports Storage Intelligence in the GA stage.

BigQuery ML now supports the following generative AI functions, which let you analyze text using a Vertex AI Gemini model. The function output includes a response that matches the type in the function name:

• AI.GENERATE
• AI.GENERATE_BOOL
• AI.GENERATE_INT
• AI.GENERATE_DOUBLE

This feature is in preview.

The following resource types are now publicly available through the analyze policy (AnalyzeIamPolicy and AnalyzeIamPolicyLongrunning) APIs.

• Compute Engine
  • compute.googleapis.com/StoragePool
• Discovery Engine
  • discoveryengine.googleapis.com/Engine

You can include pipe syntax in the SQL queries you run on the Log Analytics page. Pipe syntax supports a linear query structure designed to make your queries easier to read, write, and maintain. The pipe syntax feature is generally available (GA).

If you have enabled logging for failures of an uptime check, you can view the logs from the Uptime details page. For more information, see View details of an uptime check.

The rollout of the following extension versions and plugin versions is complete:

Extensions and plugins

• PostGIS is upgraded from 3.4.4 to 3.5.2.

To use these versions of the extensions, update your instance to [PostgreSQL version]. R20250302.00_04.

If you use a maintenance window, then the updates to the minor, extension, and plugin versions happen according to the timeframe that you set in the window. Otherwise, the updates occur within the next few weeks.

For more information on checking your maintenance version, see Self-service maintenance. To find your maintenance window or to manage maintenance updates, see Find and set maintenance windows.

Optimize log management using extractors

This feature is currently in Preview.

You can now optimize log management by creating extractors to pull specific fields from high-volume log sources. For more information, see Work with extractors.

Optimize log management using extractors

This feature is currently in Preview.

You can now optimize log management by creating extractors to pull specific fields from high-volume log sources. For more information, see Work with extractors.

Google Cloud NetApp Volumes now supports SnapMirror-based volume migration for allow-listed users. This feature lets you migrate from ONTAP-based Flex volumes to NetApp Volumes. For more information, see Volume migration.

The Secret Manager add-on for Google Kubernetes Engine (GKE) now supports the automatic rotation of secrets. You can configure the Secret Manager add-on to automatically rotate secrets so that secrets updated in Secret Manager after initial pod deployment are automatically and periodically pushed to the pod. This feature is available in Preview.

For more information, see Configure automatic rotation of secrets.

Spanner has added the PARAMETER_DEFAULT column to the INFORMATION_SCHEMA.PARAMETERS table. This column returns the default value of change stream read functions parameters.

BigQuery migration assessment now includes support for Amazon Redshift Serverless. This feature is in preview.

You can now generate structured data by using BigQuery ML's AI.GENERATE_TABLE function with Gemini 1.5 Pro, Gemini 1.5 Flash, and Gemini 2.0 Flash models. You can use the AI.GENERATE_TABLE function's output_schema argument to more easily format the model's response. The output_schema argument lets you specify a SQL schema for formatting, similar to the schema used in the CREATE TABLE statement. By creating structured output, you can more easily convert the function output into a BigQuery table.

Try this feature with the Generate structured data by using the AI.GENERATE_TABLE function tutorial.

This feature is in preview.

You can now integrate Cloud SQL for MySQL and Vertex AI (in Preview). This allows you to invoke predictions and generate vector embeddings using models hosted in Vertex AI. To use this integration, update your instance to [MySQL version].R20250304.00_01.

For more information, see Integrate Cloud SQL with Vertex AI.

GKE now provides insights and recommendations that help you identify workloads without resource requests or limits so that you can specify the resource needs for these workloads. Configuring CPU and memory requests and limits for containers is the best practice for improving reliability and performance, and is a necessary prerequisite for understanding and optimizing resource utilization by your workloads and their cost.

Migrate to Virtual Machines supports importing Arm disk images to Google Cloud. For information on operating systems supporting this feature, see Supported operating systems.

In Spanner Graph you can view a visualization of graph elements returned by a Spanner Graph query and of a Spanner Graph schema. A graph query visualization helps you understand the query results by revealing patterns, dependencies, and anomalies in the returned graph elements. A graph schema visualization helps you understand how the nodes and edges in a schema are related. For more information, see Work with Spanner Graph visualizations.

With this release, API Gateway meets the regulatory and compliance requirements for support of data residency for data at rest.

For more information, see Google Cloud Platform Services with Data residency.

VPC Service Controls (VPC-SC) integration (Preview)

API hub now integrates with VPC Service Controls, providing enhanced network security for your API hub instance provisioned in Google Cloud. Establish service perimeters to control ingress and egress traffic. For more information, see VPC Service Controls for API hub.

Data Residency Zone (DRZ) compliance

API hub is now compliant with Data Residency Zone (DRZ) C3 requirements. For more information, see API hub locations.

Terraform support for provisioning

You can now provision API hub instances programmatically using Terraform for Google Cloud within Cloud Shell, enabling infrastructure-as-code practices. For more information, see Provision API hub using Terraform.

Plugin Framework

API hub now uses a plugin framework to connect and ingest API metadata from various Google Cloud services and external sources where your APIs are managed or defined. This provides a flexible and extensible way to integrate with your existing API landscape. For more information, see Plugins overview.

API Metadata Curations

API hub introduces a curation process to transform and enrich API metadata ingested by plugins. This ensures consistency across different sources, enabling effective governance, discovery, and management of your APIs. For more information, see Curations overview.

API Supply chain graph view

Visualize and understand the dependencies within your API ecosystem with the new interactive API supply chain graph view. This directed graph allows you to explore the relationships between your APIs and API operations. For more information, see API Supply chain views.

Enhancements to the Operations entity [API only]

You can now add, edit, or delete operations for an API version even if it lacks a specification file or has an unparsable one. For more information, see Manage operations.

Deprovision an API hub instance [API only]

You can now delete an API hub instance from your Google Cloud project using the ApiHubInstance API. For more information, see Deprovision Apigee API hub.

Build Conversational Agents with Dialogflow CX (Preview)

Application Integration now simplifies the creation of conversational experiences with direct integration with Conversational Agents (Dialogflow CX). Using API triggers, you can now build intelligent chatbots and automated tools directly within your integration workflows, enhancing user interactions and automating tasks.

For more information, see Build conversational agents with Application Integration.

You can now create and use Python user-defined functions (UDFs) in BigQuery. Python UDFs support the use of additional libraries and external APIs. This feature is in preview.

You can now generate Dataframes code in BigQuery Notebooks that use BigFrames libraries. In your code generation prompt, include the word BigFrames to generate code that uses BigQuery DataFrames. This feature is in preview.

Deploying multiple containers (sidecars) to a Cloud Run job is now generally available. (GA)

Generally available: You can manage OS policy assignments across projects and zones at scale in large organizations using the OS policy orchestrator feature in VM Manager. OS policy assignment was previously available only for zonal resources in a project. For more information, see About OS Policy Orchestrator.

All processors can now extend the Maximum page limit for online and synchronous requests up to 30 pages.

To do so, enable imageless_mode in ProcessRequest.

For Custom Extractor, you will need to first request to be allowlisted for this feature by filling out the form Allowlist Request for 30 Page limit in CDE.

Salesforce CRMs: attach a CCaaS session object to a CRM record if a matching CRM record is found

For Salesforce CRMs, when you append a call or chat session to the latest open record, you have the option to attach a CCaaS session object to a CRM record if a matching CRM record is found. For more information, see Session data mapping by CRM.

New options for CRM comments when saving call recordings and chat transcripts to external storage

When you save call recordings and chat transcripts to external storage, you can control how these are referenced in the CRM record. You now have the following options:

• Add a call recording or chat transcript link as a comment in the CRM record.
• Add the call recording or chat transcript filename as a comment in the CRM record.
• Don't add any reference to the call recording or chat transcript in the CRM record.

To make comments consistent across CRM platforms, we've standardized on the following phrases in CRM comments: Chat Transcripts, Call Recordings, and Voicemails.

For more information, see Configure CRM comments.

New call type in reports: Voice Outbound (UCaaS)

We've added the Voice Outbound (UCaaS) report type to the Create Reports page for calls and chats so you can generate reports that contain this type of call. For more information, see Call and chat types.

Conditional overcapacity deflections

You can now enable conditional overcapacity deflections for calls. You can choose from a number of wait-time conditions or time-of-day conditions, and you can create a distinct deflection message for each each condition that you configure. For more information, see Configure call settings .

New post events for virtual task assistants

The following new virtual task assistant post events are available:

• Virtual task assistant joined
• Virtual task assistant left
• Virtual task assistant session variables received

The agent adapter can use the browser's postMessage() method to send events to the parent iFrame to trigger various actions in your custom CRM application. For more information, see Post events for virtual task assistants.

Bulk agent status import improvements

When you import agent statuses in bulk, the Import Statuses dialog now indicates when the upload is complete and sends you a confirmation email. For more information, see Bulk status management.

Configure a contact list destination to pass data parameters to a SIP header

You can configure a contact list destination to pass data parameters to a SIP URI when an agent uses the destination to make an outbound call or transfer a call. For more information, see Add a destination to a contact list.

View transcripts for completed chats

If you save chat transcripts in external storage, you can view them from the Completed Chats dashboard. This capability is not available in version 3.33. We expect to include it in an upcoming release.

Session metadata contains conversation IDs for virtual agents and Agent Assist

The session metadata file now contains the conversation ID for a virtual agent or for Agent Assist if either of those are involved in a session. For more information, see Sessions metadata content.

Automatic application monitoring is now generally available in GKE versions 1.28 and later. When configured on GKE clusters, this feature automatically collects key metrics with Google Cloud Managed Service for Prometheus and provides out-of-the-box dashboards for monitoring the supported workloads. Automatic application monitoring supports six new AI model servers (NVIDIA Triton, vLLM, TGI, JetStream, TorchServe and TensorFlow Serving). For more information, see Configure automatic application monitoring.

Medium Priority rule set

Google SecOps has introduced a new rule set, Medium Priority, in Applied Threat Intelligence (ATI). This rule set extends the capabilities of the ATI indicator prioritization model and expands prioritization logic to include commodity malware. For more information, see Applied Threat Intelligence priority overview.

Medium Priority rule set

Google SecOps has introduced a new rule set, Medium Priority, in Applied Threat Intelligence (ATI). This rule set extends the capabilities of the ATI indicator prioritization model and expands prioritization logic to include commodity malware. For more information, see Applied Threat Intelligence priority overview.

Memorystore for Valkey is now Generally Available (GA).

Multi-VPC support for Memorystore for Valkey is now Generally Available (GA). This functionality enables you to create Private Service Connect endpoints in multiple VPCs to connect to the same Memorystore for Valkey instance. This provides you with enhanced flexibility and resilience for your network architecture. For more information, see About multiple VPC networking.

The cross-region replication feature for Memorystore for Valkey is now Generally Available (GA). This release includes Terraform support for cross-region replication on Memorystore for Valkey.

You can now upgrade the version of your Memorystore for Valkey instance from 7.2 to 8.0. For more information, see About upgrading the Valkey version of an instance. This feature is Public Preview.

The MAC_ADDRESS_UNIVERSAL infoType detector is available in all regions. For more information about all built-in infoTypes, see InfoType detector reference.

General availability support for the following integration:

• Google Agentspace Enterprise

Pipe syntax supports a linear query structure designed to make your queries easier to read, write, and maintain. This feature is generally available (GA).

You can use a CREATE MODEL statement to create a contribution analysis model in BigQuery ML. The top_k_insights_by_apriori_support and pruning_method model options are now supported. You can use a contribution analysis model with the ML.GET_INSIGHTS function to generate insights about changes to key metrics in your multi-dimensional data. The following metric types are supported:

• Summable metric
• Summable ratio metric
• Summable by category metric

This feature is generally available (GA).

Dialogflow CX (Conversational Agents): Data store tools no longer require the use of a playbook and can be used with any agent. For information about configuring data store tools for a flow-based agent, see the data store tools documentation.

Dialogflow CX (Conversational Agents): All prebuilt agents are now GA.

Code customization for chat is generally available (GA) for VS Code and IntelliJ Gemini Code Assist. This feature provides contextually relevant code suggestions and insights in your IDE's Gemini Code Assist chat interface. Code customization for chat is available without any additional configuration required. For more information on how to use code customization for chat effectively, see Use code customization.

The following features have been added to Studio in Looker, which is available in preview:

• You can connect to Google BigQuery and Google Sheets using Owner's Credentials.
• Localized number formatting is supported.

Terraform support for using NFS solutions with SAP HANA scale-out HA deployments

While using Terraform to deploy an SAP HANA scale-out HA system on Google Cloud, you can use existing NFS solutions to share the /hana/shared and /hanabackup volumes with the worker hosts in your deployment:

• For the /hana/shared volume, use the primary_sap_hana_shared_nfs and secondary_sap_hana_shared_nfs arguments.
• For the /hanabackup volume, use the primary_sap_hana_backup_nfs and secondary_sap_hana_backup_nfs arguments.

These optional arguments are available from version 1.3.730053050 of the sap_hana_ha Terraform module provided by Google Cloud. For more information, see Terraform: SAP HANA scale-out high-availability cluster configuration guide.

New flow variable suffixes available for accessing base64-encoded message content.

There are two new read-only flow variable suffixes available for accessing message content in base64-encoded form:

• content.as.base64
• content.as.url.safe.base64

These variable suffixes can be used with the request, response, and message objects, as well as with any Message object created implicitly during API proxy execution when using the AssignMessage or ServiceCallout policies.

For more information, see Flow variables reference.

Iceberg external tables now support merge-on-read. You can query Iceberg tables with position deletes and equality deletes. This feature is generally available (GA).

On the Scheduling page, you can now view existing schedules, create new schedules, and perform other actions for data preparations, notebooks, BigQuery pipelines, and scheduled queries. For more information, see Create a pipeline schedule. This feature is generally available (GA).

You can build BigQuery pipelines (formerly workflows), composed of SQL queries or notebooks, in BigQuery Studio. You can then run these pipelines on a schedule. You can also configure notebook runtimes for a pipeline, share a pipeline, or share a pipeline link. This feature is generally available (GA).

You can now define a _CHANGE_SEQUENCE_NUMBER for BigQuery change data capture (CDC) to manage streaming UPSERT ordering for BigQuery. This feature is generally available (GA).

BigQuery now supports subqueries in row level access policies. It also includes support for BigLake managed tables and the BigQuery Storage Read API. This feature is now generally available (GA).

You can now use BigQuery Data Transfer Service for Search Ads to view Performance Max (PMax) campaign data for the following tables:

• CartDataSalesStats
• ProductAdvertised
• ProductAdvertisedDeviceStats
• ProductAdvertisedConversionActionAndDeviceStats

This feature is generally available (GA).

You can now configure the repeat frequency of BigQuery Data Transfer Service for Google Ad Manager. This option has a default of every 8 hours and a minimum of every 4 hours. This feature is generally available (GA).

You can now skip loading match tables for BigQuery Data Transfer Service for Google Ad Manager. If match tables are not needed, you can set parameter load_match_tables to FALSE. This feature is generally available (GA).

You can include data preparation tasks in BigQuery pipelines that execute your code assets in sequence at a scheduled time. This feature is in Preview.

The following resource types are now publicly available through the ExportAssets, ListAssets, BatchGetAssetsHistory, QueryAssets, Feed, and Search (SearchAllResources, SearchAllIamPolicies) APIs.

• Security Command Center Management API
  • securitycentermanagement.googleapis.com/EventThreatDetectionCustomModule
  • securitycentermanagement.googleapis.com/SecurityCenterService
  • securitycentermanagement.googleapis.com/SecurityHealthAnalyticsCustomModule

The following resource types are now publicly available through the ExportAssets, ListAssets, BatchGetAssetsHistory, QueryAssets, and Feed APIs.

• Eventarc
  • eventarc.googleapis.com/Enrollment
  • eventarc.googleapis.com/GoogleApiSource
  • eventarc.googleapis.com/MessageBus
  • eventarc.googleapis.com/Pipeline

Cloud SQL now supports Managed Connection Pooling (MCP) in Preview, which lets you scale your workloads by optimizing resource utilization for your Cloud SQL instances using pooling. To use Managed Connection Pooling, update your instance to [MySQL version].R20250302.00_04.

For more information, see Managed Connection Pooling overview.

Cloud SQL now supports Managed Connection Pooling (MCP) in Preview, which lets you scale your workloads by optimizing resource utilization for your Cloud SQL instances using pooling. To use Managed Connection Pooling, update your instance to [PostgreSQL version].R20250302.00_04.

For more information, see Managed Connection Pooling overview.

Storage batch operations for Cloud Storage is now generally available (GA). Using storage batch operations, you can perform operations on billions of Cloud Storage objects in a serverless manner. To learn more about storage batch operations, see Overview of storage batch operations.

You can now use metrics to monitor Cloud Storage FUSE performance. For more information, see Cloud Storage FUSE metrics.

Flex-start for Cloud TPU, powered by Dynamic Workload Scheduler, is available in Preview. Flex-start is a flexible and cost-effective consumption option for AI workloads. Flex-start enables you to dynamically provision TPUs for up to 7 days using the queued resources API, without long-term reservations. This option is ideal for quick experimentation, small-scale testing, dynamic inference provisioning, and model fine-tuning. For more information about Flex-start for Cloud TPU, see Request Cloud TPUs using Flex-start.

Preview: You can switch to a default runtime with GPUs by using a button in your Colab Enterprise notebook. To enable a default runtime with GPUs for your users, see Enable default runtimes with GPUs.

Support for Confidential Space on Intel CPUs (C3 machine family) with Intel TDX is now generally available.

Confidential Space now allows adding specific Linux capabilities, including CAP_SYS_ADMIN, and provides a namespaced read or write cgroup.

Dataproc Metastore federation now supports multi-regional Dataproc Metastore services.

Public preview: Google Cloud Managed Service for Apache Kafka now supports Kafka Connect. Kafka Connect provides a curated set of built-in connector plugins hosted in Connect clusters. Configure these connector plugins to create connectors that let you stream data at scale between Managed Service for Apache Kafka clusters and other systems, such as external Kafka deployments, BigQuery, Cloud Storage, or Pub/Sub. For more information, see Kafka Connect overview.

Modern charts general availability

Modern charts offers new chart styling, new default theme colors, new chart configuration options, new axis customization options, and new chart settings that give report creators greater control over how data is curated and presented to users.

This feature is now generally available and is the default for all new Looker Studio reports. Existing reports must be upgraded to use modern charts. Classic report themes are still available in the Themes panel.

Learn more about modern charts.

Looker connector enhancements

The following enhancements to the Looker connector are now generally available:

• The Looker connector can now connect to a private IP (private services access) only Looker (Google Cloud core) instance or to a private IP (Private Service Connect) Looker (Google Cloud core) instance using the Looker instance ID.

Responsive reports

You can now select between Freeform and Responsive layouts when creating a report.

• The freeform report layout is the default option. This layout is tailored for desktop screens.
• The responsive report layout scales well across many different screen sizes. Choose this layout if you expect your users to regularly view the report on a tablet or other mobile device.

Query results variables

Query result variables let you insert data directly into text elements.You can choose a cell from a table as a "query result" to insert into a text element, and Looker Studio will keep the result up to date.

• Configuration Packages: Introduces a file-based configuration package system for atomic deployments and GitOps integration. For more information, see Upload configuration package and File content.
• Development Mode: Adds a "Development Mode" to allow deletion of MDE entities and configuration packages, use with caution. For more information, see Development mode.

Spanner now supports the following GoogleSQL JSON mutator functions:

• JSON_ARRAY_APPEND()
• JSON_ARRAY_INSERT()
• JSON_REMOVE()
• JSON_SET()
• JSON_STRIP_NULLS()

Spanner now supports the following PostgreSQL JSONB mutator functions:

• jsonb_insert()
• jsonb_set()
• jsonb_set_lax()
• jsonb_strip_nulls()

Spanner also supports the following PostgreSQL JSONB operators:

• concat: jsonb || jsonb -> jsonb
• delete: jsonb - text -> jsonb

For more information, see JSON functions in GoogleSQL and Supported PostgreSQL functions.

The GoogleSQL JSON_KEYS and PostgreSQL json_object_keys functions, which extract unique JSON keys from a JSON expression, are generally available.

JSON search indexes are generally available in Spanner. This extension of Spanner's full-text index capabilities accelerates many JSON document queries, even without prior knowledge of the documents' structure. You can create search indexes over any JSON document stored in a JSON column. The JSON_CONTAINS function in GoogleSQL and the @> and <@ operators in PostgreSQL can use search indexes to determine if one document structure is contained in another. Search indexing supports JSON types in GoogleSQL-dialect databases and JSONB in PostgreSQL-dialect databases. For more information, see JSON search indexes.

You can access global Google APIs by using Private Service Connect backends that are based on cross-region internal Application Load Balancers. This feature is available in General Availability. For more information, see Access global Google APIs through backends.

Configure user preferences

The ability to manage platform time zones, date/time settings, and notifications have moved to the new User Preferences dialog, accessible from your avatar.

In addition, a new accessibility option in the User Preferences dialog lets you customize how long feedback messages remain on the screen.

For more information, see Configure user preferences.

Access Approval supports Org Lifecycle API in the GA stage.

Access Approval supports Integration Connectors in the GA stage.

Access Transparency supports Integration Connectors in the GA stage.

Access Transparency supports Org Lifecycle API in the GA stage.

The ability to disable the Invoker IAM check for Cloud Run services is now at general availability (GA).

When you create a Cloud SQL for SQL Server instance, version SQL Server 2022 Standard is now the default.

AWS token support for Confidential Space is now generally available.

You can now integrate Confidential Space with AWS resources. For more information, see Integrate AWS resources.

Instance replication is now generally available (GA).

Local codebase awareness is now available for IntelliJ Gemini Code Assist. You can now include files from your local IDE project in the prompt context by typing @ in the chat prompt box.

You can now see what files are used by IntelliJ Gemini Code Assist chat and can customize the context as needed.

In version 1.32.1-gke.1729000 and later, you can customize specific kubelet and Linux kernel parameters like sysctls and huge pages by using the nodeSystemConfig field in your GKE compute classes. Additionally, you can now specify default values for fields that are omitted in individual rules in a compute class by using the priorityDefaults field. For details, see About custom compute classes.

Finding and setting maintenance windows are now Generally Available (GA) on Memorystore for Redis Cluster.

You can now perform maintenance on a Memorystore for Valkey instance. This feature is Public Preview.

The auto-tiering feature which is previously available to allow-listed users, is now generally available. For more information, see Auto-tiering.

You can now create and manage quota rules on a NetApp Volumes volume using the Google Cloud console. For more information, see Manage quota rules.

Spanner vector index and approximate nearest neighbor (ANN) distance functions in the GoogleSQL-dialect are Generally Available. If you have a table with a large amount of vector data, you can use a vector index to accelerate similarity searches and nearest neighbor queries. Spanner now also supports the following:

• ALTER VECTOR INDEX DDL syntax
• Import and export databases that use ANN
• Use the STORING clause to store a copy of a column in the vector index to accelerate queries that filter by those columns
• Use ANN in instances smaller than one node or 1000 processing units

For more information, see Find approximate nearest neighbors, create vector indexes, and query vector embeddings.

Spanner ANN indexes are now supported in Langchain. For more information, see LangChain Quickstart for Spanner.

VPC Service Controls feature (Status: Preview): VPC Service Controls adds support for using IAM roles in ingress and egress rules to allow access to resources protected by service perimeters. This feature is available in Preview.

For more information, see Configure IAM roles in ingress and egress rules.

General availability support for the following integration:

• Storage Intelligence API

Support for a Kubernetes API connector is generally available (GA). The connector allows you to interact with Kubernetes objects in a Google Kubernetes Engine cluster. For more information, see Access Kubernetes API objects using a connector.

Availability of client IP resolution functionality with Apigee hybrid.

Client IP resolution functonality is now available with Apigee hybrid versions 1.14.0 and later.

See Client IP resolution for information.

Availability of client IP resolution functionality with Apigee hybrid.

Client IP resolution functonality is now available with Apigee hybrid versions 1.14.0 and later.

See Client IP resolution for information.

You can now enable metadata caching for SQL translation, which can significantly reduce latency for subsequent translation requests. This feature is in preview.

Cloud Workstations is available in the me-central2 region (Dammam, Saudi Arabia, Middle East). For more information, see Locations.

Dialogflow CX (Conversational Agents) data stores: Dialogflow now supports additional native and third-party data store sources as a private GA feature. For a list of data store sources, integration instructions, and the request form to be added to the allowlist, see the data stores documentation.

Parameter Manager, an extension to the Secret Manager service, is now Generally available (GA). Parameter Manager lets you store, access, and manage the lifecycle of workload parameters. You can interact with Parameter Manager using the console, gcloud CLI, REST API, and client libraries.

For information, see the Parameter Manager documentation.

You can save and manage your SQL scripts in Spanner Studio. This feature is in preview. For more information, see Saved queries overview.

General availability support for the following integration:

• Service Networking

With this release, customer data in API Gateway is now CMEK-compliant at rest. No configuration is required.

For more information, see CMEK compliance in API Gateway.

To learn more about CMEK, see Customer-managed encryption keys (CMEK).

You can now set the column granularity when you create a search index, which stores additional column information in your search index to further optimize your search query performance. This feature is in preview.

Data lineage in Cloud Composer now uses OpenLineage.

Data lineage support for a specific Airflow operator is now provided by the provider package where the operator is located. See Supported classes in the apache-airflow-providers-openlineage documentation for a list of latest supported operators.

For more information about data lineage in Cloud Composer, see Data lineage with Dataplex.

This feature is gradually rolled out. It will be available in us-west1, us-south1, europe-north1, me-west1, asia-northeast2, asia-southeast2, and africa-south1 regions. We plan to provide this feature in other regions in future releases.

Generally available: You can use instant snapshots to take in-place backups of the following types of disks:

• Hyperdisk Balanced
• Hyperdisk Balanced High Availability
• Hyperdisk Extreme

Instant snapshots are ideal for rapid data restoration only within the same location as the source disk. You can use an instant snapshot to create a new disk in under a minute. For more information, see About instant snapshots.

Generally available: You can specify a custom ephemeral internal IPv6 address when creating an instance. For more information, see Create instances that use IPv6 addresses.

Generally available: Asynchronous Replication is now generally available for Hyperdisk Balanced, Hyperdisk Balanced High Availability, and Hyperdisk Extreme disks. Asynchronous Replication provides low-RPO and low-RTO block storage replication for cross-region disaster recovery. For more information, see About Asynchronous Replication.

Custom organization policies are now available in Preview for Cloud Resource Manager. For more information, see Manage resources with custom constraints.

Generally available: You can consume reservations of VMs that have GPUs attached with your custom training jobs or prediction jobs. Reservations of Compute Engine zonal resources help you gain a high level of assurance that your jobs have the necessary resources to run. For more information, see the following:

• Use reservations with training

• Use reservations with prediction

The ability to back up and restore data on a Vertex AI Workbench instance is now generally available. For more information, see Back up and restore data on an instance.

Support for the following is available in General availability for dual-stack configurations:

• IPv6 static routes with a next hop internal passthrough Network Load Balancer (next-hop-ilb)
• IPv6 static routes with a next hop instance identified by address (next-hop-address)

For more information, see Next hops and features in the static routes overview.

API Gateway now supports Workforce Identity Federation.

Workforce Identity Federation lets you use an external identity provider (IdP) to authenticate and authorize a workforce — a group of users, such as employees, partners, and contractors — using Identity and Access Management (IAM) to access API Gateway services.

See Identity federation: products and limitations for more information.

New features added to public preview of Risk Assessment v2

This release introduces new features to the Risk Assessment v2 preview:

• Security monitoring conditions. Security monitoring conditions allow you to map resources (proxies or environments) to security profiles. Cloud Monitoring can then use this mapping to alert or create dedicated dashboards so that you can track security scores over time.
• Alerts on security monitoring conditions. Once you've created a monitoring condition, you can set up alerts using Alerting in Cloud Monitoring so that you're notified when the security scores change.

For information on monitoring conditions features and usage see monitoring conditions and alerts. For usage information and a list of all features in Risk Assessment v2, see the Risk Assessment v2 customer documentation.

Note: Rollouts of this functionality to production instances will begin within two business days and may take four or more business days to complete across all Google Cloud zones. Your instances may not have the feature available until the rollout is complete.

BigQuery ML now supports visualization of model monitoring metrics. This feature lets you use charts and graphs to analyze model monitoring function output. The following functions support metric visualization:

• ML.VALIDATE_DATA_SKEW: compute the statistics for a set of serving data, and then compare them to the statistics for the data used to train a BigQuery ML model in order to identify anomalous differences between the two data sets.
• ML.VALIDATE_DATA_DRIFT: compute and compare the statistics for two sets of serving data in order to identify anomalous differences between the two data sets.

This feature is in preview.

New services using GPUs by default will have zonal redundancy turned on. However, you can now specify GPUs with zonal redundancy or without zonal redundancy, and request quota for either of these configurations. (In Preview)

Cloud SQL read pools provide operational simplicity and scaling for your large read workloads.

Read pools provide a single endpoint in front of up to 20 read pool nodes and automatically load balance traffic.

You can scale your read pool in several ways:

• Scale in or out: scale load balancing capacity horizontally by modifying the number of read pool nodes in the read pool. Each read pool supports up to 20 read pool nodes.
• Scale up or down: scale load balancing capacity vertically by modifying the machine type associated with a read pool node. Once defined, configuration is uniformly applied across each read pool node in the read pool.

For more information, see About read pools.

Cloud SQL read pools provide operational simplicity and scaling for your large read workloads.

Read pools provide a single endpoint in front of up to 20 read pool nodes and automatically load balance traffic.

You can scale your read pool in several ways:

• Scale in or out: scale load balancing capacity horizontally by modifying the number of read pool nodes in the read pool. Each read pool supports up to 20 read pool nodes.
• Scale up or down: scale load balancing capacity vertically by modifying the machine type associated with a read pool node. Once defined, configuration is uniformly applied across each read pool node in the read pool.

For more information, see About read pools.

To send trace data to your Google Cloud project, we recommend that you use the new Telemetry API, which implements the OpenTelemetry OTLP API and provides compatibility and support for the open source ecosystem. The limits for the Telemetry API are often more generous than those for the proprietary Cloud Trace API, which you can continue to use. The Telemetry API supports VPC Service Controls. For more information about the Telemetry API, see the following documents:

• Telemetry API overview
• Migrate from the Trace exporter to the OTLP endpoint
• Quotas and limits

• DeepSeek-V3-0324, TxGemma and Sesame CSM are now available in Model Garden.
• DeepSeek-R1, V3 and V3-0324 can be deployed with H200 GPUs and improved vLLM support.
• You can deploy a model in Model Garden by using the Python SDK, gcloud CLI, or API, which are available in Preview. You can get started with the "Equivalent code" in the deploy panel in the Model Garden console.

The backups feature for the Flex service level is now generally available. For more information, see About NetApp Volumes.

Google Cloud NetApp Volumes now supports cross-region backup vaults in Preview. For more information, see Backup vaults.

The Flex service level of Google Cloud NetApp Volumes now supports custom performance in Preview, enabling independent provisioning of capacity and performance with zonal pools in selected regions. For more information, see NetApp Volumes key features.

Preview stage support for the following integration:

• Telemetry (OTLP) API

Apigee Spaces is now generally available (GA) for use in Apigee organizations.

Apigee Spaces enables identity-based isolation and grouping of API resources within an Apigee organization. With Apigee Spaces, you can have granular IAM control over access to your API proxies, shared flows, and API products.

Spaces also provide the option of resource isolation at a team level, providing a clear separation of resources associated with different teams operating within the same Apigee organization. IAM policies can be applied at the Space level, eliminating the need to manage permissions individually for every API proxy, shared flow, and API product.

Spaces are a brand new resource type with resource-level permissions. This means that Space permissions are not subject to the 64k limitation for project-level IAM conditions. Each space has its own 64k limit.

To learn more, see Apigee Spaces overview.

You can now use KLL quantile functions to efficiently compute approximate quantiles. This feature is in preview.

You can now set labels on reservations. These labels can be used to organize your reservations and for billing analysis. This feature is in preview.

The BigQuery Data Transfer Service can now transfer reporting and configuration data from Google Analytics 4 into BigQuery. This feature is in preview.

We have redesigned the Add Data dialog to guide you through loading data into BigQuery with a source-first experience and enhanced search and filtering capabilities. This feature is generally available (GA).

The following resource types are now publicly available through the ExportAssets, ListAssets, BatchGetAssetsHistory, QueryAssets, Feed, and Search (SearchAllResources, SearchAllIamPolicies) APIs.

• Binary Authorization
  • binaryauthorization.googleapis.com/Attestor
  • binaryauthorization.googleapis.com/PlatformPolicy
  • binaryauthorization.googleapis.com/Policy
• Compute Engine
  • compute.googleapis.com/InstanceSettings
• Network Security
  • networksecurity.googleapis.com/AuthorizationPolicy
  • networksecurity.googleapis.com/ClientTlsPolicy
  • networksecurity.googleapis.com/ServerTlsPolicy

Cloud SQL now lets you retain existing backups after an instance is deleted. These consist of on-demand and automatic backups created when the instance was live. For more information, see Retained backups.

Cloud SQL now lets you retain existing backups after an instance is deleted. These consist of on-demand and automatic backups created when the instance was live. For more information, see Retained backups.

You can export the transaction logs for all Cloud SQL for SQL Server instances that have point-in-time recovery (PITR) enabled and their logs stored in Cloud Storage.

Cloud SQL now lets you retain existing backups after an instance is deleted. These consist of on-demand and automatic backups created when the instance was live. For more information, see Retained backups.

Generally available: Multi-writer support for Hyperdisk Balanced High Availability disks. You can give up to 8 VMs, across two zones, simultaneous read-write access to the same disk. For more information, see Share disks between instances.

Added support for NVIDIA GB200 GPU with 570.124.06 GPU driver. This driver version has been assigned the latest, default, and R570 tags for this GPU type.

Added support for the Lustre 2.14.0 client drivers.

Added support for NVIDIA 570.124.06 GPU driver. Updated the LATEST GPU driver label to version 570.124.06 for all GPU devices. Updated the DEFAULT GPU driver label to version 570.124.06 for NVIDIA_B200 and NVIDIA_H200 GPU devices.

Add support for iRDMA devices.

Updated cos-gpu-installer to v2.4.8: Add the -skip-nvidia-smi flag to disable the execution of nvidia-smi verification during gpu driver installation.

Applied Intel patches to add iRDMA support in the Linux kernel.

Support for NVIDIA B200 GPU – Added support for the R570 driver series, including version 570.86.15. This version has been assigned the latest, default, and R570 tags.

Updated cos-gpu-installer to v2.4.7: 1.Added Support for NVIDIA B200 GPU. 2.Enabled --prepare-build-tools flag to preload GPU driver metadata for ARM64

Added support for NVIDIA GB200 GPU with 570.124.06 GPU driver. This driver version has been assigned the latest, default, and R570 tags for this GPU type.

Added support for the Lustre 2.14.0 client drivers.

Added support for NVIDIA 570.124.06 GPU driver. Updated the LATEST GPU driver label to version 570.124.06 for all GPU devices. Updated the DEFAULT GPU driver label to version 570.124.06 for NVIDIA_B200 and NVIDIA_H200 GPU devices.

Added support for NVIDIA 570.124.06 GPU driver. Updated the R570, LATEST GPU driver label to version 570.124.06 for all GPU devices. Updated the DEFAULT GPU driver label to version 570.124.06 for NVIDIA_B200 and NVIDIA_H200 GPU devices.

Added support for NVIDIA GB200 GPU with 570.124.06 GPU driver. This driver version has been assigned the latest, default, and R570 tags for this GPU type.

Added support for the Lustre 2.14.0 client drivers.

Cloud Firestore now supports multi-region nam7 United States (Central and East), which consists of regions us-central1 (Iowa) and us-east4 (Northern Virginia).

For a full list of supported locations, see Locations.

Firestore in Datastore mode now supports multi-region nam7 United States (Central and East), which consists of regions us-central1 (Iowa) and us-east4 (Northern Virginia).

For a full list of supported locations, see Locations.

The following features have been added to Studio in Looker, which is available in preview:

• The Looker connector can now connect to a private IP (private services access) only Looker (Google Cloud core) instance or to a private IP (Private Service Connect) Looker (Google Cloud core) instance using the Looker instance ID.
• The Looker connector now supports Looker export, download, and scheduling permissions.
• Looker System Activity now includes usage, historical, and performance information about Looker reports.
• You can now enable Studio in Looker on Looker instances that use Google OAuth authentication.
• The Looker connector now supports some calculated field functions.

Looker connector enhancements

The following enhancements to the Looker connector are available in Preview:

• The Looker connector can now connect to a private IP (private services access) only Looker (Google Cloud core) instance or to a private IP (Private Service Connect) Looker (Google Cloud core) instance using the Looker instance ID.
• The Looker connector now supports some calculated field functions.

Media CDN supports dynamic compression in General Availability.

After you create a Memorystore for Redis Cluster instance, you can now change the node type for the instance. For more information, see Scale an instance.

After you create a Memorystore for Valkey instance, you can now change the node type for the instance. For more information, see Scale an instance.

Preview stage support for the following integration:

• Parameter Manager

Theme enhancement for SOAR platform

The header and left hand navigation menu now fully reflect the selected theme. If you select the light theme, both the header and side menu will also appear in light mode. This might impact customers who are using the rebranding feature.

Access Context Manager now supports custom organization policies. This feature is generally available (GA). For more information, see Create custom constraints for Access Context Manager.

You can now specify, in your build config file, a custom Pub/Sub topic for build notifications. For more information, see Pub/Sub topics for build notifications.

Storage Intelligence for Cloud Storage is now generally available (GA). Storage Intelligence simplifies data management in Cloud Storage at scale by providing a unified platform for data exploration, cost optimization, security enforcement, and governance implementation. To learn more about Storage Intelligence, see Overview of Storage Intelligence.

Storage Insights datasets is now generally available (GA). Storage Insights datasets helps you get insights for your Cloud Storage resources and export the data to BigQuery. Storage Insights datasets is an exclusive feature only available through the Storage Intelligence subscription. To learn more about Storage Insights, see Overview of Storage Insights datasets.

Cross-bucket replication is now generally available (GA). You can use cross-bucket replication to copy new and updated objects asynchronously from a source bucket to a destination bucket.

Generally available: Resource-based committed use discounts (CUDs) are available for licenses of RHEL operating system images. You can purchase commitments with a 1-year plan for these licenses and receive up to 20% discounts over on-demand prices.

To learn how to purchase these commitments, see Purchase commitments for licenses. For pricing information, reach out to your Technical account manager (TAM).

The Conversational Agents console is now generally available (GA). This console combines the power of Generative AI playbooks and data stores with deterministic flows. Additional features:

• The console is now hosted at https://conversational-agents.cloud.google.com/.
• There is a new menu option for managing languages. Selecting Manage languages from the language drop-down menu at the top of the console now automatically redirects you to the Agent Settings language management section in the Deterministic Flows tab.
• The simulator now has feature parity with the Dialogflow CX console.

Conversational Agents: Conversational Agents now supports multiple tool versions in addition to playbooks and flows. See the versions and environments documentation for details.

Conversational Agents: You can now toggle a Show latency option in the Conversational Agents Console simulator to view latency per conversation turn for simulator conversations. To view latency breakdown, click the total latency value for the conversation turn. Latency values are also shown in the original response, Cloud logging, and BigQuery export.

Be aware that latency values are not recorded directly in the logs, but are represented as startTime and completeTime values in the traceBlocks field.

Conversational Agents playbooks: New model gemini-2.0-flash (Preview) is now available. For more information and supported regions, see Model support.

Conversational Agents: New Chirp 3 HD Cloud Text-to-Speech voices are now available.

Conversational Agents playbooks: Playbooks now support 38 languages. Playbook language support is displayed on the language support page. Supported languages have been tested for quality with the gemini-2.0-flash-001 and gemini-1.5-flash-002 models.

Conversational Agents playbooks: You can now enable DTMF in playbook Settings and as a conditional actions as a Preview feature. See the playbook settings and DTMF for telephony integrations pages for more information.

In GKE version 1.32.2-gke.1652000 and later, new external LoadBalancer Services use zonal Network Endpoint Group (NEG) backends by default. This applies only to new backend service-based external LoadBalancer Services. Existing LoadBalancer Services are not affected. To learn more, see Create a backend service-based external load balancer.

All GKE clusters now export four new rollup metrics by default at no additional charge. These new metrics are for monitoring GKE TPU NodePools and JobSets:

• kubernetes.io/node_pool/accelerator/times_to_recover: Distribution of recovery period durations. Each sample indicates a single recovery operation for the NodePool to recover from a downtime period. The data is sampled within 60s after the completion of NodePool recovery, and emitted within 24h. This metric does not include a sample for downtime period longer than 7 days. This metric is only applicable for GKE multi-host TPU node pools.

• kubernetes.io/jobset/times_between_interruptions: Distribution of times between the end of last interruption and beginning of current interruption for a JobSet. Each sample indicates a single duration between last and current interruption. The data is sampled within 60s after the current interruption starts, and emitted within 24h. The metric does not include a sample for duration between interruptions longer than 7 days. This metric is only applicable for JobSets running on nodes with GPU/TPU and having a single replicated job.

• kubernetes.io/jobset/times_to_recover: Distribution of recovery period durations. Each sample indicates a single recovery operation for the JobSet to recover from a downtime period. The data is sampled within 60s after the completion of JobSet recovery, and emitted within 24h. This metric does not include samples for downtime periods longer than 7 days. This metric is only applicable for JobSets running on nodes with GPU/TPU and having a single replicated job.

• kubernetes.io/jobset/uptime: Total time the JobSet has been available. The data is sampled every 60s and emitted within 24h after sampling. This metric is only applicable for JobSets running on nodes with GPU/TPU and having a single replicated job.

Starting in GKE version 1.32.1-gke.1729000, Autopilot clusters will automatically use the new Performance HPA Profile. This new profile enables faster autoscaling on CPU and Memory metrics for up to 1,000 HorizontalPodAutoscaler objects by routing autoscaling metrics through the gke-metrics-agent Daemonset. If desired, users can revert to the old autoscaling profile by disabling the Peformance HPA Profile.

Custom organization policies are now generally available for Access Context Manager and VPC Service Controls. For more information, see Manage Access Context Manager resources with custom constraints and Create custom constraints for VPC Service Controls.

VPC Service Controls now supports custom organization policies. This feature is generally available (GA). For more information, see Create custom constraints for VPC Service Controls.

You can now use repositories and workspaces in BigQuery to perform version control.

Repositories perform version control on files by using Git to record changes and manage file versions. You can use workspaces within repositories to edit the code stored in the repository.

You can have a repository use Git directly on BigQuery, or you can connect a repository to a third-party Git provider.

This feature is in preview.

You can now create remote models in BigQuery ML based on the Anthropic Claude model in Vertex AI.

Use the ML.GENERATE_TEXT function with these remote models to perform generative natural language tasks for text stored in BigQuery tables. Try this feature with the Generate text by using the ML.GENERATE_TEXT function tutorial.

You can also evaluate Claude models by using the ML.EVALUATE function.

This feature is generally available (GA).

The following resource types are now publicly available through the analyze policy (AnalyzeIamPolicy and AnalyzeIamPolicyLongrunning) APIs.

• Developer Connect
  • developerconnect.googleapis.com/Connection
  • developerconnect.googleapis.com/GitRepositoryLink

You can now use custom constraints with Organization Policy to provide more granular control over specific fields for some networksecurity and networkservices resources.

Cloud Service Mesh now supports dual-stack, extending IPv6 capability to both proxy-based Envoy and proxyless gRPC. For more information, see Configure IPv6 dual-stack for Cloud Service Mesh.

You can now use Secret Manager to securely store authentication resources with Datastream. For more information, see Use Secret Manager to store sensitive data.

Preview your data

The data source editor displays a preview of the data in your fields. This feature is now generally available.

For Autonomous Databases, you can now set up cross-region disaster recovery with Autonomous Data Guard in Google Cloud. You can create and manage peer databases, and perform switchover operations. This feature is generally available (GA).

The Risk section of the SecOps console has been updated for Security Command Center Enterprise, introducing the following features in Preview:

• Issues are the most important security risks Security Command Center Enterprise has found in your cloud environments. Sourced from Security Command Center's virtual red teaming and security graph, issues give you all the details you need to understand, triage, and remediate a risk. Explore attack path diagrams, attack exposure scores, exposed resources, related findings, and whether multiple issues exist on a primary resource, all from the one place.
• Security graph is a graph database that has cloud resources like assets, identities, apps, and data assigned to its nodes, while the edges of the graph determine the risk relationship between those resources following detection rules. When a relationship risk is discovered, the security graph generates an issue.
• Chokepoints are critical severity issues that focus on common resources or resource groups where multiple attack paths converge. Because of this focus on a common point, resolving a chokepoint can resolve other issues too, like toxic combinations.

The Risk Overview dashboard has also been updated, and a new Issues page added to the Risk section. You can navigate through different security domains in the Risk section using the tabs near the top of the page, such as All risk, Vulnerabilities, and Code.

Encrypt your data-in-use by using Confidential Computing. This feature is now available in Preview. You can enable the Confidential VM service when you create a Vertex AI Workbench instance. To get started, see Create an instance with Confidential Computing.

Vertex AI Search: Generate and return charts in answers and with follow-ups (Public preview)

The answer method can include a chart in an answer, as well as text. The chart is generated from the data in the data store. A chart is generated if there is sufficient data, and the query either asks for a chart or the answer is sufficiently complex that the method itself determines that a chart is helpful.

This feature is in public preview and is only available through the API. For more information, see Generate charts for answers.

Vertex AI Search: Return corpus images in answers and with follow-ups (Public preview)

The answer method can return images in answers, along with text.

If appropriate, one image from the data store can be returned with the answer. Citations can also include images from the data store.

This feature is restricted to queries made to unstructured data stores where the layout parser is in effect and is only available through the API. For more information, see Retrieve existing images from the data store.

Build your own Gen AI Assist is available in preview. BYOA is available in all customer engagement suite regions and offers the following:

• Foundation models
• Gemini access
• New trigger events based on agent and customer messages

Agent Assist offers Vertex extensions for Build your own assist (BYOA) in preview. Enable BYOA to access remote APIs with Vertex LLM extensions.

Performing an in-place major version upgrade of your AlloyDB cluster is generally available (GA). You can upgrade your AlloyDB cluster to any higher supported PostgreSQL version. For information about supported PostgreSQL versions, see Database version policies.

Preview: You can create regionally scoped snapshots. Setting a regional scope ensures that all snapshot data and the metadata necessary to use the snapshot are co-located within the scoped region. Regionally scoped snapshots also support additional location control by letting you restrict allowed snapshot creation and restore locations.

For more information, see Snapshot scopes.

Custom Extractor model pretrained-foundation-model-v1.4-2025-02-05 is in General Availability (GA), and has fine-tuning available in Preview for the US and EU.

From version v1.4 and later, we will use a new quota for online processing called Number of online process document pages per minute per processor_type_and_model_version. This quota will be enforced at a per-page and per-foundation model level. There will be no change to the batch processing quota.

Generally available: Migrate to Virtual Machines support for the Arm64 migration journey is now generally available. This feature lets you migrate Arm virtual machine (VM) instances from AWS and Azure cloud services to Arm VM instances on Compute Engine, and it is supported for the following operating systems:

• Debian 11 and 12
• RHEL 9
• Rocky Linux 8 and 9
• SLES 15 SP5
• Ubuntu 20.04 and 22.04

The CZECHIA_PASSPORT infoType detector is available in all regions. For more information about all built-in infoTypes, see InfoType detector reference.

Generally available: The A4 accelerator-optimized machine type is now generally available. A4 VMs are powered by NVIDIA B200 GPUs and provide up to 3x performance of previous GPU machine types for most GPU accelerated workloads. A4 is especially recommended for ML training workloads at large scales. A4 is available in the following region and zone:

• Council Bluffs, Iowa: us-central1-b

When provisioning A4 machine types, you can use Hypercompute Cluster to request capacity and create VMs or clusters. To get started see Overview of creating VMs and clusters.

Software stack updates

The following new Docker images are also released to support workloads running on your A4 GKE clusters that are deployed using Hypercompute Cluster.

• NeMo docker image: nemo25.02-gib1.0.5-A4
• MaxText docker image: jax-maxtext-gpu:jax0.5.1-cuda_dl25.02-rev1-maxtext-20150317

For more information, see AI Hypercomputer images.

You can use a query recall evaluator (Preview) to find the recall for a vector query for a given configuration, and to tune your parameters to achieve the desired vector query recall results. For more information, see Measure vector query recall.

You can now use PITR to restore a Cloud SQL for MySQL instance that isn't available.

Streamed chat responses are now available in public preview for IntelliJ and VS Code Gemini Code Assist. You can disable this feature in settings.

You can now configure and use custom commands in the inline chat menu and lightbulb menu for VS Code Gemini Code Assist. To view custom commands settings, go to Settings > Gemini Code Assist > Custom Commands.

Statistics and aggregations in UDM search using YARA-L 2.0

You can now run statistical queries on UDM events and group the results for analysis using YARA-L 2.0. You can use the statistical queries to track critical metrics, detect anomalous behavior, and analyze trends over time. For more information on how to run statistical queries on UDM events, see Statistics and aggregations in UDM search using YARA-L 2.0.

Statistics and aggregations in UDM search using YARA-L 2.0

You can now run statistical queries on UDM events and group the results for analysis using YARA-L 2.0. You can use the statistical queries to track critical metrics, detect anomalous behavior, and analyze trends over time. For more information on how to run statistical queries on UDM events, see Statistics and aggregations in UDM search using YARA-L 2.0.

Custom organization policies are now generally available for Cloud Service Mesh. For more information, see Set up custom constraints.

Cloud Infrastructure Entitlement Management (CIEM) has launched support for the following:

• Log ingestion from Amazon SQS queues.
• An alternate CIEM only feed to reduce costs.

For more information, see Configure AWS log ingestion for CIEM.

This feature is available in General Availability to the Security Command Center Enterprise tier.

The default time zone of your Spanner databases can now be set. For more information, see Set the default time zone of a database. This feature is generally available (GA).

Outbound connectivity for Private Service Connect-enabled AlloyDB clusters is generally available (GA). Enabling outbound connectivity allows secure connection between your project and an AlloyDB instance during outbound operations such as migrations or foreign data wrappers (FDW).

For Apigee organizations set up without VPC peering, you can now configure Apigee to resolve your private domains by peering your DNS zones with Apigee. See Connecting with private DNS peering zones.

You can now use EXPORT DATA statements to reverse ETL BigQuery data to Spanner. This feature is generally available (GA).

You can now create an external dataset in BigQuery that links to an existing database in Spanner. This feature is generally available (GA).

You can now use the TYPEOF function to determine the data type of an expression. This feature is generally available (GA).

Cloud Data Fusion version 6.11.0 is available in Preview.

You can view instance metrics and pipeline metrics in Cloud Monitoring and in the dashboard provided by Cloud Data Fusion. For more information, see Metrics overview and Monitor Cloud Data Fusion system, instance, and pipeline health.

You can view instance logs and pipeline logs in Cloud Logging, and in the dashboard provided by Cloud Data Fusion. For more information, see View Cloud Data Fusion logs.

When a pipeline run fails, you can retrieve detailed error information on the pipeline details page of the Cloud Data Fusion web interface.

Cloud Data Fusion classifies pipeline errors by category, reason, and message. This classification speeds up resolution and reduces the need to examine complex logs. For more information, see Retrieve error information for a failed pipeline run.

Cloud Data Fusion 6.11.0 offers high availability with reduced upgrade downtime.