GCP releases
Google Cloud releases and Terraform Google provider. New features, breaking changes, security advisories and deprecations - each summarised in plain English and updated continuously.
Tracking 329 GCP releases · Updated
- Google Cloud release notes aigcpgaengineer ·
Gemini Enterprise Agent Platform GA: Agent Gateway and Observability
Google Cloud has announced the General Availability of Agent Gateway, the networking component for the Gemini Enterprise Agent Platform, which secures and governs agent connectivity. Additionally, Agent Observability is now GA, providing visibility into agent performance and health with default-on tracing and GCS as the preferred storage for payloads. These features benefit engineers and architects managing agentic interactions and monitoring deployed agents.
feature - Google Cloud release notes aigcpgaengineer ·
Cloud SQL for PostgreSQL: AI Studio integration, minor version upgrades
Cloud SQL for PostgreSQL now integrates with Google AI Studio, enabling developers to build applications using natural language prompts for features like authentication and persistent data storage. This integration is generally available. Additionally, minor version and extension upgrades for PostgreSQL 14 through 18 are now complete, bringing new maintenance versions to production instances.
feature - Google Cloud release notes aiga ·
Gemini Enterprise adds workflow agents
Gemini Enterprise now supports workflow agents, allowing users to create and use sequences of AI and human-driven actions based on triggers. This feature is designed for complex task automation and is currently in General Availability with an allowlist. Access requires contacting a Google account manager and enabling the feature in the web app settings.
feature - Google Cloud release notes aigcpgaengineer ·
Cloud Trace GA for Generative AI Applications
Google Cloud Trace is now generally available for collecting, viewing, and analyzing multimodal prompts and responses from agentic applications using LangGraph or ADK. This enhancement allows developers to better understand and debug their generative AI applications by providing visibility into prompt and response interactions. The feature is available for all users building applications with these frameworks.
feature - Google Cloud release notes aigcpgaengineer ·
Gemini Enterprise adds custom skills and Slack integration
Gemini Enterprise is now generally available with custom skills, allowing users to create and manage reusable instructions for specific tasks. Additionally, administrators can integrate Gemini Enterprise with Slack, bringing AI-powered answers directly into user workflows. Both features aim to enhance productivity and data access through AI assistance. Custom skills require allowlist access and administrator configuration, while the Slack integration is GA for Gemini Enterprise administrators.
feature - Google Cloud release notes aigcpgaengineer ·
Gemini Enterprise Agent Platform Memory Bank and Sessions GA
Gemini Enterprise Agent Platform's Memory Bank and Sessions now support global and multi-regional endpoints in General Availability (GA). This expansion enhances data accessibility and scalability for agent development. This feature is now available for all users, with a note that Customer-Managed Encryption Keys (CMEK) are incompatible with the global endpoint.
feature - Google Cloud release notes aigcpgaengineer ·
BigQuery Autonomous Embedding Generation GA
BigQuery now offers autonomous embedding generation, automatically creating and updating an embedding column based on a source column in your tables. This feature, now generally available, simplifies AI/ML workflows by ensuring embeddings are consistently maintained as data changes. It is relevant for engineers and architects working with AI-powered data analytics.
feature - Google Cloud release notes aigcp ·
Dataflow adds NVIDIA RTX Pro 6000 GPU support
Google Cloud Dataflow now supports NVIDIA RTX Pro 6000 GPUs, enhancing its capabilities for Apache Beam pipelines. This addition is particularly beneficial for large, medium, and small model inference workloads, enabling users to leverage powerful new hardware. Users can configure these GPUs by setting the accelerator type to nvidia-rtx-pro-6000, with detailed documentation available for setup.
feature - Google Cloud release notes aigcpgapreviewdeprecationengineermedia ·
Cloud SDK 573.0.0: AI, Data Lineage, Compute Engine, and Networking updates
Cloud SDK version 573.0.0 introduces new AI capabilities, enhances data management tools like Data Lineage and Artifact Registry, and promotes several Compute Engine and Networking features to General Availability. Key additions include AI tuning job management and image streaming cache controls, alongside GA status for BigLake catalogs and Cloud Storage batch operations. This release impacts engineers and architects working with GCP services, with specific updates to Anthos Multi-Cloud and Kubernetes Engine.
breaking - Google Cloud release notes aigcppreviewengineer ·
Confidential VMs now support accelerator-optimized machine types for AI/ML
Google Cloud's Confidential VMs now support the accelerator-optimized g4-standard-48 machine type, which includes 5th Gen AMD EPYC processors and NVIDIA RTX PRO 6000 GPUs. This enhancement provides enhanced security for AI and ML workloads by encrypting data while in use. The feature is currently available in preview.
feature - Google Cloud release notes aisecuritygcpengineer ·
Google SecOps SIEM integrates Gemini Cloud Assist for feed management
Google SecOps SIEM now embeds Gemini Cloud Assist directly into its Feed Management interface, simplifying feed creation and troubleshooting. This feature aims to assist users by answering configuration questions and guiding them through setup processes. It is now available to all users within the Feed Management section, with Gemini providing recommendations that require manual application by the user.
feature - Google Cloud release notes aigaengineer ·
Gemini Enterprise integrates with ServiceNow
Gemini Enterprise now supports federation and assistant actions with ServiceNow data. This allows users to search and interact with incidents, change requests, tasks, and knowledge base articles using natural language. The integration, which is generally available, also enables actions like creating and updating incidents directly within the Gemini Enterprise app.
feature - Google Cloud release notes aipreviewengineer ·
Gemini Enterprise: Individual agent observability settings in preview
Gemini Enterprise now allows configuring observability settings for individual agents within Agent Designer. This feature, currently in Public Preview, enables monitoring specific agent metrics and trace results, offering more granular control than the previous application-level observability. This benefits engineers and architects managing complex agent deployments by providing deeper insights into agent performance.
feature - Google Cloud release notes aigcppreviewengineer ·
GCP AI Hypercomputer: Spot VM Preview Features
Google Cloud is previewing new AI Hypercomputer features for Spot VMs. These enhancements provide real-time availability, estimated uptime, preemption rates, and pricing, helping users optimize Spot VM creation for their workloads. This preview is available now and targets engineers and architects managing cloud infrastructure and costs.
feature - Google Cloud release notes aigcpengineer ·
Gemini for VS Code: Bug Fixes
Gemini for VS Code has received several bug fixes and minor enhancements. These updates aim to improve the overall stability and user experience of the Gemini extension within the VS Code IDE. The changes are generally applicable to all users of the Gemini extension for VS Code.
patch - Google Cloud release notes aigcppreviewengineer ·
Knowledge Catalog adds unstructured data profiling via Gemini
Knowledge Catalog can now profile unstructured data, such as PDFs in Cloud Storage, using Vertex AI Gemini models. This feature extracts semantic insights and relationships, enhancing understanding of diverse data sources. It is currently in Preview and accessible only through the Dataplex REST API.
feature - Google Cloud release notes aigcpgaengineer ·
BigQuery AI functions support direct ObjectRef input
BigQuery AI functions now accept ObjectRef values directly, eliminating the need for an intermediate OBJ.GET_ACCESS_URL call. This enhancement streamlines data ingestion for AI workloads and is now generally available. Developers and data scientists working with BigQuery AI will benefit from this simplified workflow.
feature - Google Cloud release notes aigaengineer ·
Gemini Enterprise mobile app now GA for Google Identity users
The Gemini Enterprise mobile app is now generally available for organizations using Google Identity, enabling users to access agents, search enterprise data, and use voice features on iOS and Android. Administrators can facilitate user access by displaying a configuration QR code on the web app homepage. For organizations using Microsoft Entra ID, mobile app access is also in GA with an allowlist.
feature - Google Cloud release notes aigcppreviewengineer ·
BigQuery adds Gemini AI assist and restores AI.KEY_DRIVERS function (Preview)
BigQuery has introduced Gemini Cloud Assist, a new feature for performance monitoring, capacity analysis, and cost optimization, currently in preview. Additionally, the AI.KEY_DRIVERS function, which identifies data segments impacting metrics, has been restored and is also available in preview. These features are primarily for data engineers and architects working with BigQuery.
feature - Google Cloud release notes aigcpengineer ·
Google Cloud Contact Center AI updates: New filters, real-time metrics, CRM links
Google Cloud's Contact Center AI (CCAI) has introduced several new features, including child queue filtering for dashboards and explores, and new real-time agent and queue metrics. Users can now link directly to CSAT scores in their CRM from CCAI dashboards, and several dashboards have been updated with new metrics and filters. These enhancements benefit contact center managers and analysts by providing deeper insights and improved workflow efficiency.
feature patch