#525 Issue
Oct. 19, 2026
Welcome to issue #524 October 12th, 2026
News
AIMigrationOfficial BlogIntroducing Google Cloud Modernize, transforming for (and with) AI - Accelerate enterprise transformation with Google Cloud Modernize and Modernization Hub, using agentic AI to assess, migrate, and refactor workloads.
Compute EngineInfrastructureNetworkingOfficial BlogIntroducing Google Cloud’s U4 compute: Enabling ultra-low latency trading - The Ultra Low Latency Solution built on the U4 machine family, now GA, provides high-frequency trading workflows with an ultra-low latency network.
JavaMCPOfficial BlogAnnouncing MCP Toolbox Java SDK v1.0: Agentic data access for the enterprise - Discover MCP Toolbox Java SDK v1.0: secure, type-safe agent orchestration for enterprise Java, with decoupled auth and AlloyDB & Spring Boot support.
AgentsGeminiOfficial BlogWorkspaceWelcome to Gemini at Work 2026: Introducing the Gemini agent - Google Cloud has introduced the Gemini agent, a universal AI assistant designed to handle complex knowledge work, write code, and integrate seamlessly into everyday business applications. The article highlights how global enterprises are leveraging Gemini's new data analytics capabilities, industry-specific workflows, and multi-model orchestration to drive unprecedented operational efficiency.
AIGemmaBring multimodal semantic search to the edge with EmbeddingGemma 2 - EmbeddingGemma 2 is a new 740M open-weight multimodal model that maps text, images, video, and audio into a unified vector space for privacy-first, on-device retrieval. Developers can easily integrate these capabilities cross-platform using MediaPipe Tasks or optimize fine-grained performance across CPU, GPU, and NPU accelerators with LiteRT.
EventOfficial BlogPublic SectorWhere mission meets moonshot: Join us at the Google Public Sector Summit 2026 - Turn moonshots into real-world outcomes at the Google Public Sector Summit on Oct. 20 in D.C. Learn how Google’s AI stack drives mission impact.
Articles, Tutorials
Infrastructure, Networking, Security, Kubernetes
SecurityNew to Google SecOps: Please, Please, Please, Let Me |> SELECT What I Want - This technical article explores how to use the SELECT pipe operator in Google SecOps to effectively reduce and refine the columns returned in your security queries. It provides practical guidance on handling nested UDM field names, formatting epoch timestamps into readable dates, and applying the DISTINCT keyword for data deduplication.
DevOps Terraformtfgcpvalidator: Catch Terraform deletion_protection failures on Google Cloud at plan time - This article introduces a new open-source tool designed to catch Terraform deletion-protection failures on Google Cloud during the plan phase rather than after apply.
Infrastructure Networking VPCIP Address Management on Google Cloud: Implementing IPAM in a Shared VPC with Terraform - This article explores Google Cloud's native IP Address Management (IPAM) capabilities, explaining how internal ranges and Cloud Number Registry automate address planning and eliminate manual bottlenecks.
App Development, Serverless, Databases, DevOps
AlloyDB Databases Official BlogAlloyDB: A unified database engine for hybrid search - Learn how AlloyDB AI eliminates complex multi-step application code, combining vector and full-text search into a single, high-performance SQL call.
Cloud SpannerSpanner — Potential impact of change stream capture modes on writes - This technical article explores how different change stream configurations in Google Cloud Spanner impact write workload performance and latency. It examines how specific value capture modes can transform efficient blind writes into read-write transactions, thereby increasing latency and the risk of lock conflicts.
Cloud Run PaywallCloud Run Can Scale Past Your Database Connection Budget - Count worker processes, pool overflow and rollout overlap before raising the instance limit.
Infrastructure StorageCost-Optimized Kafka Architecture on GCP Storage Pools - How Kafka clusters went on a storage diet when moved from a traditional storage architecture of fixed-cost disks to a Storage Pool model.
Infrastructure StorageGCP Storage Pools: Cheap Boot Disks, emptyDir Vol & FSTRIM/DISCARD - How a Storage Pool absorbed the expensive boot disks along with Trino cache, task-spill & Spark shuffle & made it almost free.
Cloud Firestore Databases NoSQLMigrating from MongoDB to Firestore Without Changing a Single Line of Code - This article explains how to migrate existing MongoDB workloads to Google Cloud's Firestore without changing a single line of code, thanks to its native MongoDB compatibility. By simply updating the connection string, engineering teams can leverage Firestore's serverless scaling, high availability, and robust security without rewriting queries or adopting new SDKs.
Big Data, Analytics, ML&AI
Apache Iceberg Cloud Spanner Data Analytics Databases Official BlogManaged Apache Iceberg at scale: How Spanner powers Lakehouse runtime catalog - Learn how the Lakehouse runtime catalog built on the Apache Iceberg REST catalog specification and powered by Spanner powers agentic workloads.
BigQuery Data Analytics GCP Experience Official Blog PartnersModernizing Unstructured Data Workflows: Alteryx Live Query meets Google Cloud BigQuery - Build code-free pipelines with Live Query + BigQuery. Use Gemini securely within your data platform to process and analyze unstructured data.
BigQuery FinOpsBigQuery Cost Optimization: The Complete FinOps Framework - This article introduces a comprehensive FinOps framework tailored specifically for Google Cloud's BigQuery, structured as a four-layer pyramid ranging from query-level optimizations to enterprise contract negotiations.
BigQuery FinOpsMaking Sense of BigQuery Pricing Models: On-Demand, Capacity, Editions (2026) - This article provides a comprehensive guide to understanding and optimizing Google Cloud's BigQuery pricing models, specifically breaking down On-Demand versus Capacity pricing and various edition tiers.
BigQueryML Machine Learning OpenTelemetryThe Zero-Cluster MLOps Blueprint: BigQuery ML, Cloud Run Jobs, and OpenTelemetry - Nightly ML scoring with no servers to patch, no clusters to size, and a distributed trace for every run.
AI Cloud Bigtable Cloud Pub/Sub Data Analytics Machine LearningPub/Sub AI Bytes: Part 5 — Zero-ETL streaming direct to database with Bigtable subscriptions - This article explores Google Cloud's new Bigtable Subscriptions for Pub/Sub, which eliminate the need for intermediary compute pipelines by enabling zero-ETL streaming directly into databases.
LookerGoogle Looker API & AI Integration Guide: Connect AI Without Admin Access. - This technical guide outlines practical methods for connecting AI tools and scripts to Google Looker without requiring full administrator access.
AI Cloud Pub/Sub DataflowEvent-Driven AI at Cloud Scale: High-Throughput Stream Classification with Pub/Sub, Dataflow, and Jev - This article explores how to build a high-throughput, event-driven AI streaming pipeline using Google Cloud Pub/Sub and Cloud Dataflow combined with TypeSafe's Jev model. By replacing traditional, resource-heavy LLMs with deterministic classification, the architecture efficiently processes millions of events per hour into typed, confidence-scored decisions.
Agents AIThe Outer Loop, Insights First: An Ambient Quality Agent That Diagnoses Your Production Agent - How do you keep improving an AI agent after the first 80%, once it serves real production traffic and silent quality regressions return HTTP 200 OK? We are open-sourcing AQuA (Ambient Quality Agent), a reference implementation that runs beside your production agent in Google Cloud, sweeps and verifies failure clusters against conversation transcripts, and diagnoses root causes anchored to the exact deployed source snapshot.
BigQuery Gemini Enterprise Agent Platform9 Mistakes to Avoid When Building Gemini Enterprise Agents - This article outlines nine critical design and architectural mistakes to avoid when building production-ready AI agents using Gemini Enterprise and BigQuery. It provides practical guidance on managing enterprise data permissions, handling identity authentication, and structuring secure tool integrations across complex business environments. The insights help developers transition from basic AI prototypes to trusted, enterprise-wide deployments with proper governance.
Gemini LLM PaywallGemini 2.5 on Vertex AI Retires October 20: What to Switch To (and What It Costs) - Google Cloud is set to retire its Gemini 2.5 models on Vertex AI by October 20, 2026, requiring developers to migrate to newer 3.x alternatives. This comprehensive guide outlines the recommended model replacements, pricing changes, and potential cost impacts for different workloads.
AI GemmaEmbeddingGemma 2: The Developer Guide - EmbeddingGemma 2 is a compact, open-source multimodal embedding model that maps text, code, images, video, and audio into a unified 768-dimensional space. Developers can use the sentence-transformers library to selectively load modular modality encoders—ranging from 270M to 740M parameters—to optimize memory usage.
AI API Generative AI Knowledge Catalog MCPSupercharge your development with the Google Developer Knowledge API ecosystem - Google is launching the Developer Knowledge API and MCP Server in public preview. This new toolset provides a canonical, machine-readable way for AI assistants and agentic platforms to search and retrieve up-to-date documentation across Firebase, Google Cloud, Android, and more.
Agents AI NetworkingGoogle Cloud’s Agentic AI Blueprint: From Design Patterns to Private Inference Networking - This article explores Google Cloud's blueprint for building production-ready agentic AI systems, moving from high-level design patterns to component selection. It details practical strategies for managing agent tools and memory, while highlighting enterprise networking requirements like private inference endpoints and secure model routing.
AI Generative AICost per task: Using cheap classifiers to route across Gemini models and thinking levels - This article explores how engineering teams can optimize the production economics of autonomous coding agents by focusing on the total cost per successful task rather than per-token API prices.
AI GPUML Drift: Next-Gen GPU AI/ML Inference at the Edge - Google AI Edge has introduced ML Drift, a high-performance, universal GPU compute framework designed to accelerate on-device AI and machine learning inference. By abstracting the complexities of hardware and low-level APIs—including OpenGL ES, OpenCL, Metal, and WebGPU—ML Drift enables developers to build real-time, interactive ML experiences, from sophisticated video effects to generative AI, across mobile and desktop platforms.
Various
Gemini Official Blog StartupsEmpowering SMBs to do more with Gemini - Millions of SMBs already trust Google AI across services like Gemini Enterprise, Google Ads, and Google Workspace Now, they can benefit from all of the agentic capabilities of Gemini.
Official Blog Retail Startups TelecommunicationsInnovation in Ireland: How Irish brands scale with Gemini Enterprise - Irish organizations spanning government agencies, established enterprise brands, and growing startups in AI and other areas are injecting agentic AI into every layer of their workflows.
Slides, Videos, Audio
GCP Bytes Podcast - #51 In this episode we discuss; Swag, Space X, Gemini Pro 4 Leaks, Le Chonk, Gemini Code Review, Waratah Super Battery, RAT Technology, Reverse Engineering, GDG, Alphabet Stock Dips, Insight get deauthorized, Darwin gets internet, Chips in space, Just try Admin, Sonnet 5.5, Haiku 5.5, Decision Models, Open AI Solves Maths, Terminator Vibes.
Releases
Anti Money Laundering AI - New minor engine version released for the commercial line of business within the v004.005 version line ( aml-commercial.default.v004.005.202609-000 ). This version extends support for the major engine version and includes no significant changes compared to the previous minor version.
Apigee Hybrid - Full notes on the release page.
BigQuery - Iceberg external tables and Iceberg managed tables now support flexible column names by default. This feature is generally available (GA). The gemini-embedding-2 model works well for embedding long strings, including multilingual and unstructured data. It supports a mix of text, images, audio, video, and PDF files. You can use this model in the AI.EMBED, AI.SIMILARITY, and AI.GENERATE_EMBEDDING functions. This model is generally available. The following table types now support flexible column names by default: Iceberg external tables and Iceberg managed tables in BigQuery Apache Iceberg tables managed by the Lakehouse runtime catalog This feature is generally available (GA). Conversational analytics in BigQuery now supports the AI.CAUSAL_EFFECT function to quantify the impact of specific interventions on time series data. This feature is in Preview. BigQuery offers TimesFM 3.0, a built-in time series forecasting model that implements Google Research's open source TimesFM model. You can use TimesFM 3.0 with the following functions: Use AI.FORECAST to perform time series forecasting. In addition to univariate forecasting, you can now perform multivariate forecasting that predicts values for multiple time series based on historical values and additional covariates. Use AI.EVALUATE to perform univariate time series evaluation. Use AI.DETECT_ANOMALIES to perform univariate time series anomaly detection. Conversational analytics in BigQuery also supports using TimesFM 3.0 with these functions. To try using the TimesFM 3.0 model for multivariate forecasting with the AI.FORECAST function, see Forecast a single time series with a TimesFM multivariate model and Forecast multiple time series with a TimesFM multivariate model. The TimesFM 3.0 model is in Preview. BigQuery offers TimesFM 3.0, a built-in time series forecasting model that implements Google Research's open source TimesFM model. You can use TimesFM 3.0 with the following functions: Use AI.FORECAST to perform time series forecasting. In addition to univariate forecasting, you can now perform multivariate forecasting that predicts values for multiple time series based on historical values and additional covariates. Use AI.EVALUATE to perform univariate time series evaluation. Use AI.DETECT_ANOMALIES to perform univariate time series anomaly detection. To try using the TimesFM 3.0 model for multivariate forecasting with the AI.FORECAST function, see Forecast a single time series with a TimesFM multivariate model and Forecast multiple time series with a TimesFM multivariate model. The TimesFM 3.0 model is in Preview. The following table types now support flexible column names by default: Iceberg external tables and Iceberg managed tables in BigQuery Apache Iceberg tables managed by the Lakehouse runtime catalog This feature is generally available (GA). You can use Gemini assistance in the BigQuery Studio SQL editor through an inline action button when you select text. For more information, see Write SQL with Gemini assistance. This feature is generally available (GA). Lakehouse now supports change data capture (CDC) ingestion from Delta Lake and Apache Iceberg tables ( Preview ). You can now synchronize CDC transaction logs (inserts, updates, deletes) from Delta Lake and Apache Iceberg tables into Lakehouse Apache Iceberg tables using Dataflow batch jobs (for Delta Lake) or batch and streaming jobs (for Apache Iceberg). For more information, see Migrate Delta Lake CDC data into Lakehouse using Dataflow and Migrate Apache Iceberg CDC data into Lakehouse using Dataflow. The BigQuery Data Engineering Agent is now available in the asia-northeast1 (Tokyo) regional endpoint and supports the gemini-3.5-flash model. For more information, see Where Gemini in BigQuery processes your data. This feature is Generally Available (GA).
Chronicle - [Spotlight Feature] Terraform and client libraries for multiple features using Chronicle API Terraform providers and Google Cloud client libraries are now available for several Google SecOps features using Chronicle API including Feeds, Parsers, Rules, Findings Refinements, Retrohunts, Dashboards, Data Tables, Watchlists, Custom Lists, Environments, Environment Groups, Data Export, SOAR Domains, and SOAR Networks. For more information, see Terraform reference and Client libraries. Terraform providers enable programmatic provisioning and management of Google SecOps resources as code, while Client Libraries enable developers to build custom integrations and automate Google SecOps workflows in their preferred programming languages.
Chronicle Security Operations - [Spotlight Feature] Terraform and client libraries for multiple features using Chronicle API Terraform providers and Google Cloud client libraries are now available for several Google SecOps features using Chronicle API including Feeds, Parsers, Rules, Findings Refinements, Retrohunts, Dashboards, Data Tables, Watchlists, Custom Lists, Environments, Environment Groups, Data Export, SOAR Domains, and SOAR Networks. For more information, see Terraform reference. Terraform providers enable programmatic provisioning and management of Google SecOps resources as code, while Client Libraries enable developers to build custom integrations and automate Google SecOps workflows in their preferred programming languages. For more information, see Client libraries. [Spotlight Feature] Terraform and client libraries for multiple features using Chronicle API Terraform providers and Google Cloud client libraries are now available for several Google SecOps features using Chronicle API including Feeds, Parsers, Rules, Findings Refinements, Retrohunts, Dashboards, Data Tables, Watchlists, Custom Lists, Environments, Environment Groups, Data Export, SOAR Domains, and SOAR Networks. For more information, see Terraform reference and Client libraries. Terraform providers enable programmatic provisioning and management of Google SecOps resources as code, while Client Libraries enable developers to build custom integrations and automate Google SecOps workflows in their preferred programming languages.
Cloud Architecture Center - Multi-agent AI system in Google Cloud and Single-agent AI system using ADK and Cloud Run: Added production security guidance for governing agent and tool interactions by using Agent Gateway, Agent Registry, Agent Identity, IAP, and IAM access policies.
Cloud Memorystore - To help protect your data and provide a baseline level of persistence, RDB snapshots are enabled by default when you create an instance using the Google Cloud console. This feature is generally available.
Cloud Monitoring - Bulk registration of discovered services and workloads to an App Hub application is generally available (GA). For more information, see Register one or more discovered services or workloads.
Cloud Run - SSH for Cloud Run services and instances is in Preview. Use this feature to establish a secure, interactive shell connection to your running instances.
Cloud SQL MySQL - Cloud SQL for MySQL now supports up to 100,000 concurrent connections for Private Service Connect enabled instances created on or after September 30, 2026 with 16 or more vCPUs. For more information, see Configure Private Service Connect. Cloud SQL for MySQL now supports the innodb_cloudsql_managed_buffer_pool_tuneup_pct database flag. Managed buffer pool automatically adjusts the InnoDB buffer pool size to reduce out-of-memory events. When instance memory usage remains at or below 70% for 10 minutes or longer after prior reductions, Cloud SQL increases the buffer pool size. You can now use the innodb_cloudsql_managed_buffer_pool_tuneup_pct flag to change this 70% threshold to a different percentage. Increase the threshold to restore the InnoDB buffer pool size when baseline usage exceeds 70%, or decrease it to preserve additional RAM allocation for memory spikes. This flag requires the Cloud SQL maintenance version MYSQL_VERSION.R20260726.00_05 or later. For more information, see Enable managed buffer pool. Cloud SQL for MySQL supports blue-green deployments ( Preview ). Blue-green deployments let you update databases—such as performing major version upgrades from MySQL 8.0 to 8.4 and staging hardware or configuration modifications—while minimizing downtime. During a blue-green deployment, Cloud SQL creates a separate staging environment (green) that mirrors your existing production environment (blue) and maintains continuous logical replication from blue to green. You can test application compatibility and performance on the green environment without affecting production traffic, and then trigger a switchover to the green environment with minimal application downtime. For more information, see About blue-green deployments in Cloud SQL. When you enable Knowledge Catalog (formerly Dataplex Universal Catalog) integration on an eligible Cloud SQL for MySQL instance created before April 18, 2026 where the integration is currently turned off, the instance now syncs schema and metadata updates to Knowledge Catalog in near real-time instead of once a day (instances created on or after April 18, 2026 have near real-time sync enabled by default). This feature is available for instances running MySQL version 8.0 or later on the new network architecture. If Knowledge Catalog integration is already enabled on the instance with once-a-day updates, you can deactivate and re-enable the integration to switch to near real-time updates. For more information, see Enable the integration of Knowledge Catalog on an existing Cloud SQL instance. MySQL 8.0.46 is now the default minor version for Cloud SQL for MySQL 8.0. For more information about minor version support in Cloud SQL for MySQL, see MySQL 8.0.
Cloud SQL Postgres - Cloud SQL for PostgreSQL now supports up to 129,024 concurrent connections for Private Service Connect enabled instances created on or after September 30, 2026 with 16 or more vCPUs. For more information, see Configure Private Service Connect. When you enable Knowledge Catalog (formerly Dataplex Universal Catalog) integration on an eligible Cloud SQL for PostgreSQL instance created before April 18, 2026 where the integration is currently turned off, the instance now syncs schema and metadata updates to Knowledge Catalog in near real-time instead of once a day (instances created on or after April 18, 2026 have near real-time sync enabled by default). This feature is available for instances running PostgreSQL version 14.0 or later on the new network architecture. If Knowledge Catalog integration is already enabled on the instance with once-a-day updates, you can deactivate and re-enable the integration to switch to near real-time updates. For more information, see Enable the integration of Knowledge Catalog on an existing Cloud SQL instance.
Compute Engine - Generally available: Z3 machine types that have 8 to 44 vCPUs can use Hyperdisk Balanced High Availability volumes to synchronously replicate data across two zones. For more information, see About Hyperdisk Balanced High Availability. Preview: External IPv6 addresses are now supported for Standard Tier. For more information, see IPv6 subnet ranges. Generally available: When you reserve compute resources for H4D instances through your account team, you can include a Hyperdisk pool in the same request or in a subsequent request. This option helps ensure that your high performance computing (HPC) workloads have the storage resources that they require close to your compute instances. For more information, see Reserve capacity through your account team. Generally available: You can sync a Compute Engine VM's clock with its host server's clock by using chrony and ptp_kvm to achieve accuracy designed to be within 1 ms. Accurate time synchronization is supported in all standard public Google Cloud regions and zones. For more information, see Configure accurate time for Compute Engine VMs.
Contact Center AI Platform - Full notes on the release page.
Dataflow - Dataflow now supports change data capture (CDC) ingestion from Delta Lake and Apache Iceberg tables into Lakehouse ( Preview ). You can use Dataflow batch jobs (for Delta Lake) or batch and streaming jobs (for Apache Iceberg) to synchronize CDC transaction logs into Lakehouse Apache Iceberg tables. For more information, see Migrate Delta Lake CDC data into borderless Lakehouse using Dataflow and Migrate Apache Iceberg CDC data into borderless Lakehouse using Dataflow.
Dataplex - Data products in Knowledge Catalog now support Looker (Google Cloud core) assets. You can package, govern, and share Looker (Google Cloud core) dashboards, dashboard elements, Looks, Explores, and LookML projects, models, and views in a data product using Google Group mirroring and IAM access controls. This feature is generally available. For more information, see About data products and Create data products.
Dataproc Serverless - Managed Service for Apache Spark (formerly Google Cloud Serverless for Apache Spark): Users may now specify short template names when creating a serverless session.
Datastream - Datastream now supports replicating SQL Server GEOMETRY and GEOGRAPHY spatial data types. For more information, see the following: Map SQL Server data types to Datastream unified types Data type mappings in BigQuery
GKE new features - GKE support for using the c4-standard-* machine types (up to 192 vCPUs) as Confidential GKE Nodes with Intel TDX is generally available. For more information, see the following pages: To use this feature with GKE, see Encrypt workload data in-use with Confidential GKE Nodes. To learn more about the feature from Compute Engine, see Confidential VM overview. To see supported configurations, including machine support, see Supported configurations and its "Machine types, CPUs, and zones" section. The microVM sandbox type is now Generally Available (GA) with GKE Sandbox in clusters that run version 1.37.0-gke.4713000 and later. MicroVM sandboxes provide hardware virtualization and isolation for untrusted workloads, AI agent runtimes, and multi-tenant environments. GKE microVM sandboxes use Kata Containers and Cloud Hypervisor to provide this isolation while maintaining a full Linux kernel that includes standard Linux features For more information, see the GKE Sandbox documentation.
Gemini - Sale of Gemini Code Assist subscriptions is ending. Starting October 9, 2026, new Gemini Code Assist Standard and Enterprise subscriptions can no longer be purchased. Existing subscriptions will continue to auto-renew through the rest of 2026, but auto-renewal will end in 2027. For more information, see Gemini Code Assist Standard and Enterprise subscriptions. For alternative AI developer tools, use Antigravity, which is available through eligible Gemini Enterprise subscriptions and through Gemini Enterprise Agent Platform.
IAM - Revoking active sessions and short-lived credentials for Workforce Identity Federation users (principals) across all clients is generally available. You can revoke workforce user sessions by using the gcloud CLI or the REST API. For more information, see Revoke Workforce Identity Federation user sessions.
Load Balancing - External IPv6 addresses are now supported for Standard Tier. For more information, see IPv6 subnet ranges. This feature is in Preview.
Looker - A full notes on the release page.
Network Intelligence Center - Flow Analyzer displays packet drops, allowing you to analyze packet loss in your traffic flows. For more information, see Display flows in packet drop mode.
Network Tiers - External IPv6 addresses are now supported for Standard Tier in Preview. For more information, see IPv6 subnet ranges.
Security Command Center - Artifact guard with CI/CD integration is deprecated. It will be shut down on October 30, 2026. Using gcloud CLI, the securitycenter REST API, or client libraries, you can configure Pub/Sub and BigQuery continuous export to receive notifications when findings are deleted from Security Command Center due to data retention policies. For more information, see the following: Enable finding notifications for Pub/Sub Stream findings to BigQuery for analysis Enable real-time notifications Creating and managing Notification Configs
VPC Service Controls - VPC Service Controls feature (Status: Preview ): Support for automated remediation suggestions is available in the VPC Service Controls violation analyzer. Remediation suggestions analyze access denial events and generate actionable, narrowly scoped configuration changes to resolve perimeter denials and adhere to the principle of least privilege. The remediation engine can automatically propose ingress rules, egress rules, VPC accessible services updates, and existing or new context-aware access levels. For more information, see Resolve access denials with remediation suggestions.
Virtual Private Cloud - Preview: External IPv6 addresses are now supported for Standard Tier. For more information, see IPv6 subnet ranges. General Availability: VPC Flow Logs supports TCP connection logging for VM instances and serverless endpoints. Connection logging captures outbound TCP connection attempts and acknowledged inbound TCP connections. You can enable this feature in a VPC Flow Logs configuration for a subnet (through the Network Management API), a VPC network, or an organization. Service producers can let service consumers establish more connections to a Private Service Connect endpoint or backend by configuring the target published service (service attachment) to use an additional NAT IP address. For more information, see Maximum connections.