Google Cloud Platform Newsletter

Check Archive for older issues

News

Anthropic’s Claude 3 Opus and tool use are generally available on Vertex AI - Anthropic’s Claude 3 Opus, the most capable and intelligent model in the Claude 3 family, is now generally available on Vertex AI. It excels at complex tasks like in-depth analysis, research, and task automation.

Easily stream data from AWS Kinesis to Google Cloud with Pub/Sub import topics - Pub/Sub import topics enable streaming ingestion into BigQuery from external sources, with the first supported external source being Amazon Kinesis Data Streams. Import topics provide a simplified way to ingest data from Amazon Kinesis Data Streams directly into Pub/Sub, reducing the complexity of setting up data pipelines between clouds. Once the connection is established, Amazon Kinesis producers can be gradually migrated to Pub/Sub publishers. Data from Amazon Kinesis Data Streams can be routed to BigQuery using BigQuery subscriptions, and Pub/Sub autoscales to adapt to changes in the Amazon Kinesis data stream.

Introducing Cloud SQL extended support for MySQL and PostgreSQL end-of-life versions - Cloud SQL now offers extended support for end-of-life MySQL and PostgreSQL versions, ensuring business continuity and security while organizations prepare for major version upgrades. Starting February 1, 2025, Cloud SQL instances running unsupported versions will automatically enter extended support, receiving security fixes, bug fixes, and SLA coverage for three years. Extended support is a paid service with pricing details to be announced by August 15, 2024. Customers can opt out of extended support at any time and upgrade to a supported version.

Transform Your Agency’s Security Posture with a New Offer from Workspace - Google Workspace offers advanced threat detection, zero-trust architecture, and continuous vulnerability patching to protect public sector agencies' data. Public sector customers can get special pricing for Workspace Enterprise Plus, Assured Controls Plus, Chrome Enterprise Premium, and training and migration assistance for a limited time. A SCuBA Secure Configuration Baselines in Google Workspace Workshop will be held on June 13th in Washington D.C. to help security directors and IT managers understand how Google Workspace can meet CISA Benchmarks.

What’s new for the Google Cloud global front end for web delivery and protection - The Google Cloud global front end solution helps customers deliver and protect internet-facing web services using the same technologies, infrastructure, and teams that Google uses for its own web services. It consists of the Cloud External Global Application Load Balancer, Cloud CDN, and Cloud Armor. Recent enhancements include Service Extension callouts for web data plane programmability, private origin access over the Internet with App Connector, custom error responses, load balancing for AI workloads, Cloud Armor Enterprise for premium web and DDoS protection, granular Adaptive Protection ML models for Layer 7 DDoS Defense, Graph QL API Protection, UI controls for dynamic compression with gzip and Brotli, Internet observability with Catchpoint, and a Global Front End CI/CD Automation Toolkit and reference guide.

Articles, Tutorials

Infrastructure, Networking, Security, Kubernetes

Building the cloud-native broadcast media supply chain with Google Cloud - Google Cloud is leading the transformation of the media supply chain by partnering with key media companies and ISVs to build cloud-native solutions. This shift enables broadcasters to streamline operations, reduce costs, and deliver more engaging content to global audiences. With a focus on openness, efficiency, and AI integration, Google Cloud is helping media companies unlock the full potential of the cloud for their future success.

Cloud FinOps and GKE: How Generali optimized spend with Google Cloud Consulting - Generali Switzerland engaged Google Cloud Consulting to optimize its GKE clusters and workloads, resulting in annual savings of over €260,000, a 41% reduction in GKE spend. The optimization focused on four key areas: workload rightsizing, demand-based autoscaling, bin packing, and discount coverage. Generali achieved significant savings by rightsizing requests, fixing autoscaling issues, migrating to cost-optimized machines, and maximizing discount coverage.

FQDN filtering in Cloud Next Generation Firewall: A complete guide - The new FQDN feature in Cloud Next Generation Firewall (NGFW) allows users to specify domain names (like www.google.com) in firewall rules, simplifying management and eliminating the need to track IP addresses. FQDNs offer improved reliability, ease of use, and enhanced security by reducing the impact of IP address changes and mitigating DNS spoofing attacks. With FQDN objects, users can create more flexible and precise firewall rules, enhancing network security while simplifying management.

Cloud CISO Perspectives: What the past year tells us about our cybersecurity future - The past year in cybersecurity has been marked by several trends, including fewer consequences for threat actors, accelerated offensive innovation by attackers, and the evolution of ransomware from data theft to extortion and potentially even harassment. Boards of directors are also engaging more in cybersecurity, and private enterprises and public organizations are cooperating more to advance security practices. Defenders are detecting attacks sooner, and more organizations are finding attackers on their systems on their own. However, attackers are also innovating, using living-off-the-land techniques, developing custom malware, and targeting consumer and local infrastructure.

5 myths about platform engineering: what it is and what it isn’t - Platform engineering is a relatively new approach to software delivery that aims to reduce friction and cognitive overload for developers by abstracting away the complexity of modern software systems. It involves creating an internal developer platform that provides developer self-service through golden paths, codifying DevOps practices into software, and taking a holistic approach to automation. Platform engineering is not just advanced DevOps or automation, and it is not a fad but a response to the growing complexity of modern software systems.

How PrimaKu Slashed GCP Costs with GKE Clusters - PrimaKu, a pioneer in child development applications in Indonesia, shares their journey of slashing Google Cloud Platform (GCP) costs by implementing Google Kubernetes Engine (GKE) clusters. By optimizing resource usage and leveraging GKE's automated resource management, efficient resource allocation, and cost-efficient scaling, PrimaKu achieved a remarkable 62% reduction in total GCP costs within a few months.

A Complete Guide to Deploy Main Services in GCP with Terraform - Introduction to Terraform and GCP.

Deploying Airbyte on GKE using Helm - This blog post provides a step-by-step tutorial on how to successfully deploy Airbyte on Google Kubernetes Engine (GKE) using the official Helm chart. The outcome of this tutorial is a GKE cluster that can automatically scale, Airbyte deployed on Kubernetes, an external Postgres database for Airbyte metadata, and an external location for Airbyte logs. The tutorial covers topics such as service account creation, cluster creation, namespace creation, load balancer configuration, Cloud SQL setup, log bucket creation, deploying Airbyte using Helm, accessing the web server, and pod disruption budgets.

Landing Zone Deployment (Google Cloud Adoption Series) - Step-by-step guidance for how to actually deploy our LZ, either using Cloud Setup “Click-Ops” in the console, or with Terraform.

How a Google Cloud Administrator can enforce security with Organization Hierarchy, Organizational Policies, and IAM Policies - Google Cloud administrators can enforce security using Organization Hierarchy, Organizational Policies, and IAM Policies. The organization hierarchy enforces the principle of least privilege, while organizational policies restrict how resources are used and are inherited from the parent node. IAM policies control who can do what on which resource, and pre-defined roles are recommended by Google Cloud. These building blocks provide a solid secure foundation on which to layer on more advanced security services.

App Development, Serverless, Databases, DevOps

Cloud Run: the fastest way to get your AI applications to production - Cloud Run offers a straightforward way to deliver AI-powered applications to production, allowing developers to focus on their application logic without worrying about the underlying infrastructure or scaling. It provides features like fast prototyping, observability, rapid innovation with concurrent revisions, and connectivity to cloud databases for integrating with enterprise data. Custom domains and multi-regional deployments with a global external load balancer ensure a reliable user experience.

Shipping Fast with FastAPI and Cloud Run - This article demonstrates how to quickly build and deploy a backend API server using FastAPI and deploy it to Google Cloud Run. The API interacts with Google's Firestore database using the Discovery API and Firestore Client libraries. The code is available on an open-source repository, and the deployment process involves creating a Docker image, pushing it to Artifact Registry, and deploying it to Cloud Run. Testing and verification steps are also provided.

Beam moves to Google Cloud to drive efficacy of government safety net programs - Beam, a digital social safety net platform, migrated its infrastructure from AWS to Google Cloud to enhance efficiency and reduce costs. By adopting Google Cloud services like Cloud Run and AlloyDB for PostgreSQL, Beam streamlined its operations, improved scalability, and achieved significant cost savings. The migration process was smooth, with no downtime or data loss, and the platform's performance has since improved, enabling Beam to effectively distribute funds across various program areas. Additionally, Beam appreciates Google Cloud's prioritization of security, developer-friendly tools, innovation enablement, and social and environmental responsibility.

Five ways to save on Compute Engine costs - Compute Engine offers various ways to reduce costs, including optimizing infrastructure and taking advantage of discounts. The first step is to inspect your current spending structure using the Google Cloud console's billing panel to identify unnecessary resources. Automated recommendations from Recommendation Hub can suggest actions like changing machine types or scaling down machines. Reviewing disk types and optimizing disk utilization can also lead to significant savings. Additionally, committing to using Committed Use Discounts (CUDs) can provide substantial discounts for long-term virtual machine usage.

How DZ BANK improved developer productivity with Cloud Workstations - DZ BANK, Germany's second-largest bank, partnered with Google Cloud to enhance developer productivity and security. They utilized Cloud Workstations to standardize development environments, enabling faster onboarding and improved collaboration. Preconfigured base images simplified infrastructure management, while persistent disks and Docker-in-Docker support enhanced flexibility. DZ BANK also implemented custom Docker images, project-specific customizations, and an order CI pipeline for efficient image management and deployment.

Architecting a blockchain startup with Google Cloud - Google Cloud offers a comprehensive suite of products and services to help blockchain startups build, deploy, and manage their applications quickly, securely, and efficiently. These services include highly scalable infrastructure, high-performance networking, blockchain data analytics, and support for Web3 development. Google Cloud also provides financial support, learning opportunities, and access to advanced technology through its Web3 startup program.

Multi-tenancy in Spanner: How many instances should I have? - Spanner instances are fundamental building blocks in Cloud Spanner, acting as containers for databases that share processing units (PUs) and storage. The optimal number of instances depends on factors like resource utilization, workload isolation, and operational ease. A tiered approach to instances can be helpful, categorizing applications into tiers based on their criticality and resource requirements.

Cloud Trace spanning PubSub - Tracing across PubSub (or other asynchronous action).

Automatically remove old docker container image using Artifact Registry cleanup policies - This article summarises how to use that cleanup policies to automatically remove old container images in Artifact Registry.

Google Cloud Artifact Registry Goes Limitless with Generic Format Support - Google Cloud Artifact Registry now supports generic format storage, allowing users to store and share any type of artifact, from simple text files to compressed archives. This eliminates the need for using different tools for storing and sharing non-standard artifacts, configuration files, compressed files, binaries, and media files in build and deployment tools.

Big Data, Analytics, ML&AI

Build a hybrid data processing footprint using Dataproc on Google Distributed Cloud - Dataproc on Google Distributed Cloud enables organizations to modernize their data lake infrastructure while maintaining regulatory compliance by processing sensitive on-prem data locally and moving the rest to the cloud. It supports full local execution of Spark jobs, allowing for aggregation and anonymization of sensitive data before uploading it to the data lake on the cloud. This hybrid data processing approach ensures data residency requirements are met while still enabling comprehensive data analysis and integration with Google Cloud Data Analytics services.

Accelerating CDC insights with Dataflow and BigQuery - This post covers how to use BigQuery’s new CDC capability in Dataflow along with the new Dataflow at-least-once streaming mode to simplify your CDC pipeline and reduce costs.

Data platform from scratch on GCP - Solvimon's bespoke analytics experience.

A Tale of Two Functions : Function calling in Gemini - Gemini, a large language model, introduces a new feature called "Function calling." This feature allows users to integrate their own functions or APIs with the model, enabling it to provide more specific and accurate responses to queries. The model can determine which functions to call based on the user's prompt and can even make multiple function calls in parallel. This opens up new possibilities for building powerful applications that combine the strengths of natural language processing with custom business logic.

Terraforming Dataform - Dataform 101, Part 2: Provisioning with Least Privilege Access Control.

Combining ANY_VALUE with HAVING in BigQuery - In this article, Constantin Lungu demonstrates a unique use case for the ANY_VALUE function in BigQuery. He explains how to combine ANY_VALUE with the HAVING clause to find orders that contain a single value from a specified list. By utilizing the arbitrary value retrieval capability of ANY_VALUE and filtering the results with HAVING, Lungu efficiently identifies orders consisting of exactly one item, such as grapes or oranges. While alternative methods exist, this approach offers a less obvious but effective solution.

Dynamically Generating MERGE Queries with BigQuery’s Procedural Language - BigQuery's scripting language offers powerful features for automating repetitive tasks, including dynamically creating and updating queries. Utilizing this capability to generate MERGE queries on the fly, it eliminates the need for manual maintenance and significantly reduces the time required for syncing data from our landing zone to target tables.

Grounding Gemini with Web Search results in LangChain4j - The latest release of LangChain4j (version 0.31) added the capability of grounding large language models with results from web searches. This allows the LLM to find relevant information about the query from web searches, which will likely include up-to-date information that the model won’t have seen during its training. The article provides a step-by-step guide on how to implement this feature using Google Custom Search Engine and Gemini 1.5 Flash model. The author also discusses some limitations and potential improvements of this approach.

Vertex AI's Grounding with Google Search: how to use it and why - Vertex AI's Grounding with Google Search helps improve the factuality and up-to-date information of large language models (LLMs) by grounding their responses on trusted Google Search world knowledge and public facts. It addresses the limitations of LLMs, such as hallucinations, staleness, lack of citation, and limited relation to private data. With Grounding, LLMs can provide more reliable and trustworthy responses, especially for questions that require recent or factual information.

Hugging face and Open AI Models in AlloyDB Omni - In this post, I'll show you how to register AI models hosted by third-party providers, such as OpenAI and Hugging Face, in AlloyDB Omni. We'll cover the process of enabling the Secret Manager API, generating API keys, creating secrets in Google Secret Manager, and registering models using the google_ml extension. Examples of registering the OpenAI gpt-4o model and the Hugging Face Mistral-7B-Instruct-v0.3 model are provided, along with sample queries to test the models.

Slides, Videos, Audio

Kubernetes Podcast - #226 A Decade of Kubernetes Contribution.

Security Podcast - #174 How to Measure and Improve Your Cloud Incident Response Readiness: A New Framework.

Releases

Agent Assist - Agent Assist now offers Summarization with custom sections as a GA feature.

Anthos Config Management - 1.18.1. Upgraded bundled Helm version from v3.14.3 to v3.14.4 to pick up vulnerability fixes. Upgraded the Open Telemetry image from v0.91.0-gke.9 to v0.99.0-gke.1 to pick up vulnerability fixes. Fixed an issue where Config Sync installation would fail when using a private registry with a specified port in the image URL.

Google Distributed Cloud Bare Metal - 1.16. Release 1.16.9 Google Distributed Cloud for bare metal 1.16.9 is now available for download. Fixes: The following container image security vulnerabilities have been fixed in 1.16.9: High-severity container vulnerabilities: CVE-2023-39323 CVE-2023-45285 Medium-severity container vulnerabilities: CVE-2023-39318 CVE-2023-39319 CVE-2024-2961 CVE-2023-39326 CVE-2024-28085 Low-severity container vulnerabilities: CVE-2023-5363 CVE-2023-6246 CVE-2023-6779 CVE-2023-6780. Known issues: For information about the latest known issues, see Google Distributed Cloud for bare metal known issues in the Troubleshooting section. 1.16 & 1.28 & 1.29. Security bulletin (all minor versions) A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

Anthos clusters on bare metal - 1.16. Release 1.16.9 Google Distributed Cloud for bare metal 1.16.9 is now available for download. Fixes: The following container image security vulnerabilities have been fixed in 1.16.9: High-severity container vulnerabilities: CVE-2023-39323 CVE-2023-45285 Medium-severity container vulnerabilities: CVE-2023-39318 CVE-2023-39319 CVE-2024-2961 CVE-2023-39326 CVE-2024-28085 Low-severity container vulnerabilities: CVE-2023-5363 CVE-2023-6246 CVE-2023-6779 CVE-2023-6780. Known issues: For information about the latest known issues, see Google Distributed Cloud for bare metal known issues in the Troubleshooting section. 1.16 & 1.28 & 1.29. Security bulletin (all minor versions) A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

GKE on AWS - A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

Anthos clusters on Azure - A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

GDCV for VMware - A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

Apigee Advanced API Security - On May 29, 2024 we released a new version of Advanced API Security NOTE: Rollouts of this feature are ongoing and will take multiple days to complete across all Google Cloud zones. Preview release of Shadow API Discovery This release introduces Shadow API Discovery in preview.

Apigee X - On May 29, 2024 we released an updated version of Apigee. Preview release of API Management features in Gemini Code Assist: generative AI API spec creation with enterprise context and Apigee policy code explanation.

Apigee Hybrid - 1.12.0-hotfix.1. ANNOUNCEMENT hybrid 1.12.0-hotfix.1 On May 28, 2024 we released an updated version of the Apigee hybrid software, 1.12.0-hotfix.1. Bug ID Description 340889560 Added csi to the apigee-logger SCC.

Cloud Architecture Center - Design an optimal storage strategy for your cloud workload: Added information about the Regional service tier of Filestore. (New guide) Build an ML vision analytics solution with Dataflow and Cloud Vision API: Deploy a Dataflow pipeline to process large-scale image files with Cloud Vision.

Cloud Asset Inventory - The following resource types are now publicly available through the Analyze IAM Policies APIs (AnalyzeIamPolicy and AnalyzeIamPolicyLongrunning).

Bare Metal Solution - You can now order Bare Metal Solution storage and Partner Interconnect resources on a 1 month commitment term.

BigQuery - You can now use IAM conditions to control access to BigQuery resources. You can now define a _CHANGE_SEQUENCE_NUMBER for BigQuery change data capture (CDC) to manage streaming UPSERT ordering for BigQuery. The maximum number of partitions per partitioned table limit has changed from 4,000 to 10,000. The following Generative AI features are now in preview: Creating remote models based on the Vertex AI gemini-1.5-flash foundation model.

Chronicle - Default parsers have changed, see release notes URL for more information.

Chronicle SOAR - Release 6.3.4 is now in General Availability. Release 6.3.5 is currently in Preview. Trying to set an SLA definition that is too similar to an existing one results in an incorrect error message (ID #00289305). Tags not showing as expected in the Search page (ID #50691614). All Environments is not supported when importing networks from CSV (ID #00276371). Action All CVE Entity filter is not working (ID #51310124). Subject Entity Search Filters are not working properly (ID #50841312). Case actions - generate report has missing content (ID #50620576).

Cloud Composer - Cloud Composer 2 now supports data lineage for environments that have CMEK enabled. Cloud Composer 2.8.1 images are available: composer-2.8.1-airflow-2.7.3 (default) composer-2.8.1-airflow-2.6.3.

Compute Engine - Creating a larger (>90 vCPUs) C3D standard-lssd or highmem-lssd VM results in an error message.

Contact Center AI Platform - Web SDK 2.2 is released For more information, see Web SDK changelog. Version 3.16 is released All release notes published on this date are part of version 3.16. End co-browse sessions using the Apps API The Apps API has the following new endpoint that lets you end a co-browse session using an external session ID: POST /apps/api/v1/cobrowse_sessions/{external_session_id}/end. The agent adapter generates co-browse events The agent adapter generates events during co-browse sessions. The Next UI is supported in the ServiceNow integration The Next UI experience is supported in the ServiceNow CRM integration. Fixed an issue that prevented agents from selecting their next status to exit a campaign when the current call is concluded. Fixed the problem of the created_at field being missing from add_started_activity.

Dataflow - You can now use Metrics Explorer to find individual DoFns that cause latencies in streaming jobs.

Dataform - Dataform Core includeDependentAssertions and dependOnDependencyAssertions parameters for adding assertions as dependencies are available.

Dataplex - Dataplex automatic data quality supports the following capabilities: Email notifications to alert people about the status and results of a data quality job Data quality scores that indicate the percentage of rules that passed API support for rule recommendations based on data profiling scans For more information, see Use auto data quality and Auto data quality overview.

Dataproc Metastore - Dataproc Metastore services can now enable deletion-protection to prevent the accidental removal of new or existing services.

Dataproc Serverless - New Dataproc Serverless for Spark runtime versions: 1.1.63 1.2.7 2.0.71 2.1.50 2.2.7. Dataproc Serverless for Spark: Subminor version 2.1.50 is the last release of runtime version 2.1, which will no longer be supported and will not receive new releases. Dataproc Serverless for Spark: Removed Spark data lineage support for runtime version 1.2. Dataproc Serverless for Spark: Enabled Spark checkpoint (spark.checkpoint.compress) and RDD (spark.rdd.compress) compression in the latest 1.2 and 2.2 runtime versions.

Dialogflow - Dialogflow CX: You can now integrate with Soul Machines to create 3-D avatars.

Document AI - Model pretrained-foundation-model-v1.2-2024-05-10 is available for custom extractor.

Integration Connectors - Custom connectors now support the following additional features: Edit connector Versions Logo File upload for OpenAPI specification.

Google Kubernetes Engine - GKE cluster versions have been updated. A new vulnerability (CVE-2024-4323) has been discovered in Fluent Bit that could result in remote code execution.

Live Stream API - You can now create video on demand (VOD) clips from a live stream.

Cloud Logging - Ops Agent version 2.47.0 introduces support for Compute Engine VMs that are running Ubuntu 24.04 LTS (Noble Numbat).

Marketplace Partners - We've added a new field, cancellation_reason, on the Entitlement resource that provides context around why an entitlement was cancelled.

Memorystore for Redis Cluster - Added support for Deletion protection for Memorystore for Redis Cluster.

Cloud Monitoring - Ops Agent version 2.47.0 introduces support for Compute Engine VMs that are running Ubuntu 24.04 LTS (Noble Numbat). Announcing new Open Telemetry samples that show how to instrument your Python and Node.js applications to collect metrics, logs, and traces: Python example Node.js example For general instrumentation information and recommendations, and for links to other samples, see: Instrumentation and observability overview Choose an instrumentation approach.

NetApp - The Standard in Preview service level is now called Flex and is generally available. The volume replication feature for the Flex service level is now generally available. The Flex service level now supports zone-redundant storage pools (in Preview). NetApp Volumes now supports auto-tiering (in Preview).

Network Connectivity Center - Preset topologies are now available in public preview.

reCAPTCHA Enterprise - reCAPTCHA Enterprise Mobile SDK v18.5.1 is now available for iOS. reCAPTCHA Enterprise Mobile SDK v18.5.1 is now available for Android. reCAPTCHA SMS toll fraud protection is now available in Preview.

SAP Solutions - Version 1.7 of the ABAP SDK for Google Cloud is generally available (GA).

Cloud Spanner - Spanner now supports the protocol buffer data type in GoogleSQL. A monthly digest of client library updates from across the Cloud SDK. Spanner now supports the following new columns in the SPANNER_SYS query statistics table: AVG_MEMORY_PEAK_USAGE_BYTES AVG_MEMORY_USAGE_PERCENTAGE AVG_QUERY_PLAN_CREATION_TIME_SECS AVG_FILESYSTEM_DELAY_SECS AVG_REMOTE_SERVER_CALLS AVG_ROWS_SPOOLED.

Cloud SQL MySQL - Cloud SQL for MySQL major versions that have reached community end-of-life (EOL) will receive extended support starting on February 1, 2025.

Cloud SQL Postgres - Cloud SQL for PostgreSQL major versions that have reached community end-of-life (EOL) will receive extended support starting on February 1, 2025.

Cloud SQL SQL Server - Cloud SQL for SQL Server now supports storage of point-in-time recovery (PITR) transaction logs in Cloud Storage.

Cloud Storage - Cloud Storage FUSE now offers the following features: You now have the option to disable authentication on custom endpoints using the --anonymous-access flag.

Cloud Trace - You can now search a trace for keywords. Announcing new Open Telemetry samples that show how to instrument your Python and Node.js applications to collect metrics, logs, and traces: Python example Node.js example For general instrumentation information and recommendations, and for links to other samples, see: Instrumentation and observability overview Choose an instrumentation approach.

Cloud Translation - For adaptive translations, when you use the API, you can include up to five reference sentence pairs in a request instead of specifying a dataset.

Vertex AI - Model Monitoring v2 is in Preview, which centralizes model monitoring configuration and visualization on a model version and enables monitoring models being served outside of Vertex AI. Vector Search sparse embeddings and hybrid search in Public preview Vector Search supports sparse embeddings and hybrid search in Public preview.

Workflows - Support for the following connectors is generally available (GA): AI platform training and prediction Application Integration, Batch API ,BigQuery Data Transfe,r Cloud Resource Manager, Google Forms, Google Sheets ,Transcoder API. The maximum number of concurrent workflow executions has increased from 5,000 to 7,500.

Workstation - The Code-OSS preconfigured base image uses version 1.89.1. Cloud Workstations is available in the southamerica-east1 region (Osasco, São Paulo, Brazil, South America). Cloud Workstations is available in the us-east5 region (Columbus, Ohio, North America).