Google Cloud Platform Newsletter

Check Archive for older issues

News

An open approach to security: Announcing Google Unified Security Recommended - Introducing Google Unified Security Recommended, a new program that establishes strategic partnerships with market-leading security solutions.

Accelerating discovery at the speed of cloud: What’s New for HPC at Google Cloud for SC25 - At SC25, come learn how Google Cloud is democratizing HPC for startups, academics, labs, and enterprise teams alike.

Expanding support for AI developers on Hugging Face - Growing partnership will reduce Hugging Face model download times for Google Cloud users and offer native support for TPUs while enhancing security.

Announcing BigQuery-managed AI functions for better SQL - BigQuery managed AI functions let you use generative AI for common analytical tasks within SQL queries, without complex prompt tuning or new tools.

Do you detect a pattern? BigQuery’s new MATCH_RECOGNIZE function can! - The new MATCH_RECOGNIZE feature in BigQuery allows you to perform complex pattern matching on your data directly within your SQL queries.

Introducing the Emerging Threats Center in Google Security Operations - To help organizations learn if they’ve been affected by vulnerabilities, we’re introducing the Emerging Threats Center in Google Security Operations.

Talk with and trust your data using Looker’s Conversational Analytics, now GA - Looker Conversational Analytics, now GA, transforms how you get answers about data, letting you use natural language to accelerate data discovery.

Introducing Agent Sandbox: Strong guardrails for agentic AI on Kubernetes and GKE - Agent Sandbox is a new Kubernetes primitive designed specifically for agent code execution and computer use.

Supporting Viksit Bharat: Announcing our newest AI investments in India - We’re investing in powerful local tools in India to foster a diverse ecosystem and ensure our platform delivers controls for compliance and AI sovereignty.

Achieve better AI-powered code reviews using new memory capabilities on Gemini Code Assist - Today, we’re releasing a new memory capability for Gemini Code Assist on GitHub for both enterprises and individual developers. Now, you can create a dynamic, evolving memory of your team's coding standards, style, and best practices.

N4D now GA: Gain up to 3.5x price-performance for scale-out workloads - With up to 96 vCPUs, N4D VMs with AMD EPYC Turin and Titanium (now GA) deliver excellent price-performance and flexibility for everyday workloads.

Articles, Tutorials

Infrastructure, Networking, Security, Kubernetes

Time Travel Triage: An Introduction to Time Travel Debugging using a .NET Process Hollowing Case Study - The basics of WinDbg and Time Travel Debugging necessary to start incorporating it into your analysis.

Running high-scale reinforcement learning (RL) for LLMs on GKE - Learn about Google Cloud's full-stack, integrated approach, from custom TPU hardware to the GKE orchestration layer – and shares how you can solve the hybrid, high-stakes demands of RL at scale.

Google Cloud Networking under the hood: How Protective ReRoute increases resilience - Google’s Protective ReRoute empowers communicating endpoints to detect a failure and intelligently re-steer traffic to a healthy, parallel path.

GKE: From containers to agents, the unified platform for every modern workload - At KubeCon North America, we announced major advancements to take both open-source Kubernetes and Google Kubernetes Engine (GKE) into the next decade.

No Place Like Localhost: Unauthenticated Remote Access via Triofox Vulnerability CVE-2025-12480 - An unauthenticated access vulnerability in Gladinet's Triofox platform, exploited by the threat actor UNC6485.

App Development, Serverless, Databases, DevOps

A new top score: Advancing Text-to-SQL on the BIRD benchmark - Every database is unique. With this finetuning recipe, Gemini will write accurate SQL for your natural language queries.

Simplicity Meets Speed: Expanding One-Time Migrations to Heterogeneous Workloads in Google DMS - Database Migration Service (DMS) expands its One-Time Migration capability to support heterogeneous migrations, offering a simplified "dump and load" workflow. This new mode removes the need for Change Data Capture (CDC), eliminating prerequisites like enabling ARCHIVELOG on Oracle or configuring change tables on SQL Server.

Building a Firestore Database CLI with Rust - A Rust-based CLI tool to interact with a remote Google Cloud Firestore database, leveraging the Gemini CLI and native Rust libraries.

Waze keeps traffic flowing with 1M+ real-time reads per second on Memorystore - To scale successfully, Waze needed a shared, low-latency solution that could handle user sessions in real time and at global scale. Memorystore for Redis Cluster made that possible.

Zeotap's big win: 46% TCO reduction and enhanced real-time performance with Bigtable - Zeotap’s ScyllaDB-based infrastructure faced scaling challenges, leading them to Bigtable and a 46% reduction in Total Cost of Ownership (TCO).

Big Data, Analytics, ML&AI

A Google Engineer’s Take on a Common Spark Problem — And How We’re Fixing It - Google Cloud is developing Autotuning for Serverless Apache Spark to automate performance optimization, addressing common issues like executor sizing, shuffle partitioning, and join strategies.

How We Designed the Data Infrastructure for One of France’s Largest Sports/Sneaker Retailers - A French sports retailer faced data management challenges due to isolated silos. By integrating MuleSoft, custom APIs, and Google Cloud serverless resources, a scalable data pipeline was built.

Concurrent DAG Execution Capacity with Shared IP Subnets in Google Cloud Composer - Google Cloud Composer's concurrent DAG execution capacity is limited by shared IP subnet ranges, impacting the number of worker nodes and parallel tasks. Resolving this involves larger IP ranges or separate VPCs for each environment to ensure long-term growth and stability.

BigQuery under the hood: How Google brought embeddings to analytics - We reflect on the past two years since launching vector search in BigQuery, with a focus on scale, simplicity, and cost performance.

BigQuery Join Order: Let the Optimizer Win - How to write join-friendly SQL that’s fast, cheap, and easy to maintain — without micromanaging the plan.

Zeotap: How BigQuery ML and vector search help customers build their own AI models - Zeotap uses BigQuery, including BigQuery ML and Vector Search, to solve the end-to-end lookalike problem in its Zeotap CIP platform.

RAG just got much easier with File Search Tool in Gemini API - The article discusses Google's new File Search Tool in the Gemini API, which simplifies Retrieval Augmented Generation (RAG) by providing a fully managed system for grounding Gemini with user data.

How Lightricks trains video diffusion models at scale with JAX on TPU - From Bottleneck to Breakthrough: How Lightricks Trains Video Diffusion Models at Scale with JAX.

Building a Multi-Agent ADK/A2A Malware Analysis Agent: For Fun and Profit - Written together with my friend and colleague Eran Tal, who has architected the entire system and built the code!

Building A2A Agents with the GO Agent Development Kit (ADK) - Native GO Agent Coding with the GO ADK and the A2A protocol.

Building PocketGuide: A Multi-Agent Platform for Personalized Audio Tours - Built for the Google Cloud Run AI Agents Hackathon. Runs entirely on serverless infrastructure.

Technical guide: Four steps for startups to build multi-agent systems - In this post, we’ll show you a systematic, operations-driven roadmap for navigating this new landscape, using one of our projects to provide concrete examples for the concepts laid out in the official startups technical guide: AI agents.

Building a Personalized Ad Generation Assistant with Google Cloud’s Agent Development Kit - Google Cloud's Agent Development Kit (ADK) enables the creation of a "Product Catalog Ad Generation Agent" that uses generative AI to automate personalized video ad creation. The agent leverages product metadata and user feedback to generate storylines, images, and audio, significantly reducing ad production time.

Building Cloud AI Copilots: Multi-Agent Systems on Google Cloud Run with ADK - Building collaborative AI agents with Gemini 2.5, ADK, and the A2A Protocol.

Various

Securing the mission: Google Public Sector’s CMMC Level 2 certification and commitment to national security - Google Public Sector achieves CMMC Level 2 certification, reinforcing commitment to DoD missions with secure cloud, AI, and data services for national security.

I did it! I Became a Google Developer Expert! - A journey of becoming a Google Developer Expert.

Slides, Videos, Audio

Security Podcast - #251 Beyond Fancy Scripts: Can AI Red Teaming Find Truly Novel Attacks?

Agent Factory Recap: Cracking Open an Open Model - Learn the rigorous, end-to-end process of building and training open-source agentic models—from defining agent data and combining SFT/RL to the critical role of intense evaluation.

Releases

AlloyDB - Feature AlloyDB AI native vector search accelerator is now generally available (GA). Feature You can now use Dataplex Universal Catalog to discover, search, and manage your AlloyDB resources, including clusters, databases, and tables. Feature Improve query performance for partitioned tables and complex queries using parallel execution for the SELECT portion of an INSERT INTO .. Feature AlloyDB AI introduces auto embedding generation in Preview. Feature AlloyDB now supports PostgreSQL 17 for migrating from Cloud SQL for PostgreSQL to AlloyDB for PostgreSQL using your Cloud SQL for PostgreSQL backup.

Anthos clusters on VMware - Google Distributed Cloud (software only) for VMware 1.31.1100-gke.40 is available for download. The following issues were fixed in 1.33.200-gke.70: Fixed an issue that gkectl update blocks updating a non-HA non-advanced cluster to HA advanced cluster Fixed vulnerabilities listed in Vulnerability fixes. The following issues were fixed in 1.31.1100-gke.40: Fixed vulnerabilities listed in Vulnerability fixes. Google Distributed Cloud (software only) for VMware 1.33.200-gke.70 is available for download.

API Gateway - Announcement On November 12, 2025, we released a new version of API Gateway. v1. Feature API Gateway now supports OpenAPI version 3.0.

Apigee UI - Announcement On November 12, 2025, we released an updated version of the Apigee UI. Fixed Bug ID Description 455584175 Fixed a performance issue with Debug session UI Fixed an issue where performance of the Debug session was severely degraded when loading a Debug session with a moderate number of transactions.

App Hub - Feature App Hub has expanded its support for Google Cloud resources. Feature Management projects are now Generally Available (GA) to enable application management in folders.

AppEngine Standard Python3 - Feature Support for Python 3.14 runtime is in Preview.

BigQuery - Feature The BigQuery Overview page is now your hub for discovering tutorials, features, and resources to help you get the most out of BigQuery. Feature You can now use the interactive SQL translator, the translation API, and the batch SQL translator to translate the following SQL dialects into GoogleSQL: Apache Impala SQL GoogleSQL (BigQuery) Impala SQL translation can be used to migrate Cloudera and Apache Hadoop SQL workloads that use Impala as a query engine. Feature You can now use custom constraints with an Organization Policy to provide more granular control over specific fields for BigQuery dataset resources. Feature You can aggregate and deduplicate table data with Gemini assistance in your BigQuery data preparations. Feature Partitioning is now available for BigLake tables for Apache Iceberg in BigQuery. Feature BigQuery ML now supports the TimesFM 2.5 time series foundational model. Feature BigQuery ML now offers the AI.DETECT_ANOMALIES function.

CDN - Feature Cloud CDN lets you add custom code to the request processing path of global external Application Load Balancers by using Service Extensions edge extensions (pre-cache) and traffic extensions (post-cache).

Chronicle - Feature Raw log search enhancements Google SecOps now includes enhancements to raw log search to boost usability, performance, and data analysis: New filtering options: Filter raw log results by their parsing status or by one or more log sources. Announcement New parser documentation now available New parser documentation is available to help you ingest and normalize logs from the following sources: Collect Absolute Secure Endpoint logs Collect AIDE (Advanced Intrusion Detection Environment) logs Collect Akamai Enterprise Application Access logs Collect Apache Hadoop logs Collect Armis Vulnerabilities logs Collect Array Networks SSL VPN logs Collect Aruba IPS logs Collect Atlassian Confluence logs Collect Cisco AMP for Endpoints logs Collect Cisco APIC logs Collect Cisco Application Centric Infrastructure (ACI) logs Collect Cisco CallManager logs Collect Cisco CloudLock CASB logs Collect Cisco DNA Center Platform logs Collect Cisco eStreamer logs Collect Cribl Stream logs Collect CrowdStrike FileVantage logs Collect CrowdStrike IDP Services logs Collect Cynet 360 AutoXDR logs Collect Digital Shadows SearchLight logs Collect Duo Telephony logs Collect Edgio WAF logs Collect Elastic Auditbeat logs Collect Elastic Packet Beats logs Collect Elasticsearch logs Collect Entrust nShield HSM audit logs Collect Imperva Advanced Bot Protection logs Collect Imperva Attack Analytics logs Collect Imperva Audit Trail logs Collect Imperva CEF logs Collect Imperva Data Risk Analytics (DRA) logs Collect Imperva Database logs Collect Imperva FlexProtect logs Collect Imperva SecureSphere Management logs Collect Kiteworks (formally Accellion) logs Collect Proofpoint Emerging Threats Pro IOC logs Collect ServiceNow audit logs Collect Team Cymru Scout Threat Intelligence data Collect URLScan IO logs Collect Uptycs EDR logs Collect VanDyke VShell SFTP logs Collect Zendesk CRM logs Collect ZeroFox Platform logs. Feature Enhance threat visibility and detection with Emerging Threats The new Emerging Threats page provides AI-powered threat intelligence to help you understand how current threat campaigns might affect your organization. Feature Use the Triage Agent to investigate alerts You can now use Triage Agent, an AI-powered investigation assistant, to analyze alerts in Google SecOps. Feature Nested if You can now use if statements in both the outcome and events sections and also within the then else clauses of another if statement.

Chronicle SOAR - Announcement Release 6.3.68 is being rolled out to the first phase of regions as listed here. Feature Change views per alert The Alert Overview page now includes a new option to select a specific view for each alert. Announcement Release 6.3.67 is now available for all regions.

Colab - Feature The default latest Python version is now 3.12.

Cloud Composer - Feature You can now start Gemini Cloud Assist investigations for failed Airflow tasks.

Compute Engine - Feature Preview: Future reservation requests in calendar mode support reserving capacity for a3-megagpu-8g and a3-highgpu-8g machine types. Issue The Windows guest agent identifies administrator accounts and groups using string matching. Feature Generally available: Two new features that provide better observability for Compute Engine reservations and deeper insights into their capacity usage and costs: A new consumedReservation field in the VM instance details showing the full resource name of the consumed reservation, providing reservation consumption status on VMs and better visibility for resource management and troubleshooting.

Config Connector - Announcement Config Connector version 1.138.0 is now available. Feature New Beta Resources (Direct Reconciler): BackupDRBackupVault OrgPolicyCustomConstraint. Feature New Alpha Resources (Direct Reconciler): FirestoreBackupSchedule FirestoreDocument. Change Reconciliation Improvements: Improved Normalization logic for OrgPolicy, RunJob, TagsTagBinding, and VertexAIIndex resources. Fixed Bug Fixes: Fixed format validation issue in the DataflowFlexTemplateJob direct controller when the spec.subnetworkRef.external field contains full URL.

Contact Center AI Platform - Announcement Google Cloud CCaaS 3.42 pre-release notes Here are the pre-release notes for Google Cloud CCaaS version 3.42. Feature Alvaria Advanced List Management integration with outbound dialer You can now integrate Alvaria Advanced List Management (ALM) with the outbound dialer. Feature New standalone queue settings page Google Cloud CCaaS now includes a standalone Queue Menu Settings page for each queue in your instance. Feature Restrict the transfer of email sessions to a queue You can now configure queues to prevent them from receiving email session transfers. Feature Email signatures You can now define standardized signatures for all outgoing emails sent from a specific queue. Feature Outbound calling improvements We've implemented the following improvements to outbound calling: Show dialpad before starting the call: For CRM users, the Settings > Operation Management > CRM Operations section contains a new Show dialpad before starting the call checkbox. Feature Virtual agent alias You can assign a public-facing alias to each of your virtual agents. Feature Automatic redaction You can now automatically identify and redact sensitive data from chat conversations, both in real time and in transcripts. Feature New chat platform API endpoint for getting media files You can use the new chat platform API endpoint to get the media file that an agent sends to an end-user during a chat session. Feature Virtual task assistant support in the chat platform API The chat platform API now provides improved support for virtual task assistants. Feature Web SDK: chat check-in Users of the web SDK can use chat check-in to help ensure that end-users are present and ready to engage before connecting them to a human agent. Fixed We addressed the following issues in this release: Fixed an issue where Agent Assist generated incomplete or inaccurate chat summaries.

Dataform - Feature You can use custom constraints with Organization Policy to provide more granular control over specific fields for the CompilationResult, ReleaseConfig, WorkflowConfig, WorkflowInvocation, and Workspace resources.

Dataproc Serverless - Announcement New Serverless for Apache Spark runtime versions: 1.2.62 2.2.62 2.3.15. Feature Serverless for Apache Spark: Added the dataproc.artifacts.remove property, which lets users remove default artifacts, such as spark-bigquery-connector, iceberg, and delta-lake from a Serverless for Apache Spark runtime.

Document AI - Feature Automated schema extraction for custom extractor processors is in Preview.

Cloud Endpoints - Feature Cloud Endpoints now supports OpenAPI version 3.0.

Eventarc - Feature Eventarc Advanced support for using access policies to control access when publishing event messages to an Eventarc Advanced bus is available in Preview.

Cloud NGFW - Feature You can create a firewall endpoint that supports jumbo frames up to 8,500 bytes in size.

Cloud Functions - Feature Support for Python 3.14 runtime is in Preview.

Gemini - VS Code Gemini Code Assist 2.57.0 & VS Code Gemini Code Assist 2.58.0. Issue Known issue in VS Code 2.57 and later The Gemini Code Assist 2.57 and later releases for VS Code have a known issue that prevents the extension from connecting to the server if you are using a gnc c runtime version that's earlier than 2.36. Feature Chat performance metrics Gain deeper insights into Gemini for Google Cloud's chat performance with metrics for chat acceptance rate and lines of chat accepted, including interactions in Agent mode and non-agentic chat. Feature Code customization available in agent mode and Gemini CLI (Preview) Code customization is now supported when using Gemini CLI and agent mode in Gemini Code Assist. IntelliJ Gemini Code Assist 1.38.0. Feature Inline diff is generally available (GA) for IntelliJ Inline diff is generally available for IntelliJ Gemini Code Assist. Feature Persistent memory for Gemini Code Assist on GitHub (Preview) Gemini Code Assist on GitHub now supports persistent memory, which stores your previous interactions with Gemini Code Assist on GitHub so that it has context during your future interactions.

KMS - Feature The Cloud KMS Encryption metrics dashboard is available in Preview.

GKE new features - Feature The N4D machine family is now Generally Available (GA) for Standard and Autopilot mode.

Looker - Feature Conversational Analytics is now generally available for Looker instances that are on Looker 25.18 or later. Feature You can now schedule a daily, zero-downtime export for Looker (Google Cloud core) instances.

Migration Center - Announcement The discovery client 6.3.9 is available with new features and bug fixes. Feature Preview: Added support for discovery of the following assets from your AWS account: Amazon DynamoDB Amazon Route53 NAT Gateway For more information, see Discover assets on AWS. Fixed Fixed an issue that prevented the mcdc discover ls command from displaying IP addresses. Fixed Fixed an issue that caused the VMware vCenter status to incorrectly display the previous status during a new collection, such as warning or error. Fixed Fixed an issue that caused the collection of data from Linux machines to become unresponsive on systems running JBoss EAP.

Network Intelligence Center - Feature The following insight types and subtypes are available in Network Analyzer: Recommender insight subtypes: ROUTE_NEXT_HOP_FORWARDING_RULE_IP_MISMATCH ROUTE_NEXT_HOP_FORWARDING_RULE_TYPE_INVALID ROUTE_NEXT_HOP_ILB_VIP_NOT_ASSIGNED ROUTE_NEXT_HOP_IP_NOT_ASSIGNED ROUTE_NEXT_HOP_VM_NON_PRIMARY_IP ROUTE_NEXT_HOP_VM_WRONG_NETWORK ROUTE_NEXT_HOP_VPN_TUNNEL_NOT_ESTABLISHED NCC_DYNAMIC_ROUTE_FULLY_SHADOWED Cloud Logging insight types: ROUTE_INVALID_NEXT_HOP_FORWARDING_RULE_IP_MISMATCH ROUTE_INVALID_NEXT_HOP_FORWARDING_RULE_TYPE_INVALID ROUTE_INVALID_NEXT_HOP_ILB_VIP_NOT_ASSIGNED ROUTE_INVALID_NEXT_HOP_IP_NOT_ASSIGNED ROUTE_INVALID_NEXT_HOP_VM_NON_PRIMARY_IP ROUTE_INVALID_NEXT_HOP_VM_WRONG_NETWORK ROUTE_INVALID_NEXT_HOP_VPN_TUNNEL_NOT_ESTABLISHED DYNAMIC_ROUTE_SHADOWED_FULLY_SHADOWED_BY_NCC_SUBNET_ROUTE NCC_DYNAMIC_ROUTE_SHADOWED_FULLY_SHADOWED_BY_PEERING_SUBNET_ROUTE NCC_DYNAMIC_ROUTE_SHADOWED_FULLY_SHADOWED_BY_STATIC_ROUTE NCC_DYNAMIC_ROUTE_SHADOWED_FULLY_SHADOWED_BY_SUBNET_ROUTE For more information, see Routes with an invalid next hop and Shadowed dynamic routes. Change Network Analyzer no longer generates the following insight types and subtypes: Recommender insight subtypes: ROUTE_NEXT_HOP_ILB_BACKEND_IP_FORWARDING_DISABLED DYNAMIC_ROUTE_PARTIALLY_SHADOWED PEERING_DYNAMIC_ROUTE_PARTIALLY_SHADOWED Cloud Logging insight types: ROUTE_INVALID_NEXT_HOP_ILB_BACKEND_IP_FORWARDING_DISABLED DYNAMIC_ROUTE_SHADOWED_PARTIALLY_SHADOWED_BY_SUBNET_ROUTE DYNAMIC_ROUTE_SHADOWED_PARTIALLY_SHADOWED_BY_PEERING_SUBNET_ROUTE DYNAMIC_ROUTE_SHADOWED_PARTIALLY_SHADOWED_BY_STATIC_ROUTE DYNAMIC_ROUTE_SHADOWED_PARTIALLY_SHADOWED_BY_PEERING_STATIC_ROUTE.

Cloud Run - Feature You can deploy services to Cloud Run using a Compose file (Preview). Feature Support for Python 3.14 runtime is in Preview. Feature Cloud Run source deployment supports Ubuntu 24 LTS base images (Preview). Feature You can set a task timeout up to 168 hours (7 days) for Cloud Run jobs.

Security Command Center - Change Several features and updates have been made available to Security Command Center in a federated identity environment: Exporting findings to a CSV file. Feature Model Armor is available in the following regions: europe-west1 (Belgium) europe-west2 (London) europe-west3 (Frankfurt) asia-south1 (Mumbai) For more information, see Locations for the Model Armor API. Feature Data Security Posture Management (DSPM) has been released to General Availability for the Security Command Center Enterprise tier.

Service Extensions - Feature Edge extensions help you manipulate request headers early in the processing lifecycle of global external Application Load Balancers to influence caching and routing decisions. Feature Regional external and internal Application Load Balancers support route and traffic extensions using plugins.

Cloud Spanner - Feature Spanner Studio now supports several new export options for your query results. Feature Spanner now automatically provides recommendations to apply schema design best practices to your databases.

Cloud Speech-to-Text - Feature Speech-to-Text has just launched chirp_3 in public preview for the regions asia-south1, europe-west2, europe-west3 and northamerica-northeast1.

Cloud SQL MySQL - Feature Cloud SQL for MySQL now supports read pool autoscaling (GA), which helps you more easily manage your application's workload needs. Feature Cloud SQL for MySQL now lets you have more control over the number of results that are returned when you perform an ANN vector search with filters. Feature Cloud SQL for MySQL now supports minor version 8.0.44. Feature Cloud SQL now provides brute-force detection and protection for your Cloud SQL instances.

Cloud SQL Postgres - Feature Cloud SQL for PostgreSQL now supports read pool autoscaling (GA), which helps you more easily manage your application's workload needs.

Cloud Storage - Feature You can now use the Google Cloud console to relocate buckets. Feature You can now enable Autoclass for buckets with hierarchical namespace enabled.

Cloud Text-to-Speech - Feature Chirp 3: HD voices now support speech synthesis in the following languages: bg-BG, cs-CZ, et-EE, el-GR, he-IL, hr-HR, hu-HU, lt-LT, lv-LV, ro-RO, sk-SK, sl-SI, and sr-RS.

VMware Engine - Announcement VMware Engine ve2 nodes are now available in the Tokyo, Japan region (asia-northeast1-a).

Virtual Private Cloud - Feature Dynamic Private Service Connect interfaces are available in General Availability. Feature Service producers can accept or reject connections from individual Private Service Connect endpoints.