Google Cloud Updates

March 2025

BigQuery: Version Control & Code Assist

Repositories and workspaces are now available for code management with Git. This means you can use Git directly to manage your projects more efficiently, or integrate external providers to make your workflows even more flexible. Additionally, the innovative Gemini Cloud Assist now offers the ability to generate both SQL and Python code in BigQuery. Both features are currently available in preview.

Google Workspace Gemini

The “Take Notes for Me” and transcription features are now available in additional languages:

• German
• Italian
• Japanese
• Korean
• Portuguese
• Spanish
• French

More information can be found here and here.

February 2025

Generative AI on Vertex AI

The code adaptation for Gemini Code Assist Enterprise now supports repositories hosted on the following platforms:

• GitHub Enterprise Cloud
• GitHub Enterprise
• GitLab
• GitLab Enterprise
• Bitbucket Cloud
• Bitbucket Data Center

BigQuery Updates

BigQuery data preparation now provides context-aware recommendations for Gemini join operations. This feature is available in preview. Additionally, the BigQuery Data Transfer Service can now transfer data from MySQL and PostgreSQL. This feature is also available in preview.

New Gemini 2.0 Pro and Gemini 2.0 Flash-Lite Models available

Two new models from the Gemini 2.0 family are now available to users:

• Gemini 2.0 Pro: Google’s most powerful model for coding and world knowledge, featuring a 2M token context window. Gemini 2.0 Pro is available as an experimental model in Vertex AI.
• Gemini 2.0 Flash-Lite: Google’s fastest and most cost-efficient Flash model. Gemini 2.0 Flash-Lite is available as a preview model in Vertex AI.

Security Command Center Updates

Protect your AI applications with Model Armor: Model Armor is a Google Cloud service that enables you to apply content safety and security controls to LLM prompts and responses. It helps mitigate risks such as sensitive data leaks, prompt injection, and offensive content. Learn more.

December 2024

Gemini for Google Workspace now with multilingual support

• Gemini for Google Workspace is now available in 7 additional languages, including German.
• Users can now interact with Gemini in their native language in the side panel of Docs, Sheets, Drive and Gmail.
• Support for Slides will be added in the coming months.

read more

Generative AI in BigQuery (General available)

BigQuery ML has been enhanced with new features that make it possible to perform generative AI tasks directly in BigQuery:

• Remote models: You can create remote models based on the Vertex AI models gemini-1.5-flash and gemini-1.5-pro.
• Text generation: The ML.GENERATE_TEXT function can be used to perform generative tasks on text data in BigQuery tables.
• Multimodal processing: ML.GENERATE_TEXT enables the processing of image, video, audio, PDF and text content from BigQuery object tables, e.g. for audio transcription or document classification.

read more

New machine learning capabilities in BigQuery (General available)

BigQuery ML now offers advanced language model customization and evaluation capabilities:

• Supervised Tuning: Remote models based on Vertex AI Gemini 1.5 flash or Gemini 1.5 pro can now be optimized using supervised tuning.
• Model Evaluation: The ML.EVALUATE function can be used to evaluate Vertex AI language models. Pre-trained PaLM and Gemini models are supported, as well as optimized Gemini models.
• Instructions for using the new features can be found in the how-to topic “Customize an LLM by using supervised fine tuning” and in the tutorial “Use tuning and evaluation to improve model performance”.

read more

Gemini-enhanced SQL translation in BigQuery

BigQuery now offers the following Gemini-enhanced SQL translation features:

• Interactive translation mode (GA): Customization of translated Google SQL queries using Gemini.
• AI suggestions for batch translations (Preview): Generation of suggestions based on a Gemini-based configuration YAML file.
• Text-based explanations (preview): After an interactive SQL translation, a Gemini-generated explanation with a summary of the translated SQL query can be requested.

Generative AI on Vertex AI

• Batch predictions for Llama models (MaaS) available in preview.
• Batch predictions for Gemini generally available.

Supported models are Gemini 1.0 Pro, Gemini 1.5 Pro, and Gemini 1.5 Flash. More information.

October 2024

BigQuery Migration Assessment for Oracle (Preview Feature)

The new BigQuery Migration Assessment for Oracle helps you assess the complexity of migrating data from your Oracle data warehouse to BigQuery. This feature is already available for other source systems and now also for Oracle. More on this

Gemini in BigQuery (General available)

The following Gemini features in BigQuery are now generally available to every organization and user. More on this

• Data Insights: Gain insights into your data quickly and easily.
• Data Canvas: Visualize and analyze your data interactively.
• SQL and Python Code Support: Use SQL generation tools, generate and explain
• SQL queries, and generate Python code.
• Partitioning and Clustering Recommendations: Get recommendations to optimize your table structure.

Workload management enhancements (General available)

Workload management in BigQuery has been improved and now offers the following advantages. Read more

• Instant scaling: The autoscaler now scales up instantly
• More precise scaling: The autoscaler now scales more precisely
• Smaller scaling increments: The autoscaler scales to the nearest multiple of 50 slots instead of 100 slots
• More flexible capacity management: You can now purchase capacity commitments, set baseline slots, and set the maximum number of autoscaling slots in increments of 50 slots
• Faster capacity reduction: If at least one minute has passed since the last capacity increase, you can now reduce capacity without resetting the one-minute minimum waiting period. This allows for multiple consecutive reductions without delay.

Multimodal input with function calls for Gemini (Preview)

Gemini 1.5 Pro and Gemini 1.5 Flash now support multimodal input with function calls. This feature is currently in preview. Read more

The imagen 3 image generation models (imagen-3.0-generate-001 and the low-latency version imagen-3.0-fast-generate-001) are now generally available to licensed users. More information

These models offer the following additional features:

• Additional aspect ratios: 1:1, 3:4, 4:3, 9:16, 16:9
• Digital watermark (SynthID): enabled by default
• Watermark Verification: Enables you to verify the authenticity of images
• Customizable security features: Security and people/faces settings

New Gemini features in Looker (Public Preview)

• Create custom Looker visualizations: The Visualization Wizard helps you generate custom formatting options for Looker visualizations using natural language
• LookML generation: Use Gemini support to generate suggested LookML code in response to natural language input. Click the “coding guidance” icon in the Looker IDE for Gemini assistance with creating dimensions, dimension groups, and measures in your LookML project

Learn more

July 2024

BigQuery Migration Assessment Features

The BigQuery Migration Assessment now offers new features: automatic creation of datasets for storing results, real-time view of the assessment report during execution, and more detailed information as well as error reports on the assessment details page. Learn more.

BigQuery Continuous Queries Available in Preview

BigQuery introduces "Continuous Queries," a new feature that allows continuously running SQL queries. These queries analyze, process, and perform machine learning on incoming data in BigQuery in real-time. The results can be replicated into Pub/Sub topics, Bigtable instances, or other BigQuery tables. Continuous Queries are suitable for tasks such as transforming incoming data, applying real-time ML insights, building automated data-driven pipelines, and replicating real-time events to downstream systems. Learn more.

Vertex AI Gemini 1.5 Pro Context Caching

Context Caching is now available for Gemini 1.5 Pro. This feature allows you to reduce the cost of requests that contain repeated high-token-count content. Instead of sending these contents with every request, they are stored once in the cache and can be referenced as needed. This is particularly useful for applications such as chatbots with extensive instructions, repeated analyses of large video files, or recurring queries of large document sets. Learn more.

Cloud Billing FinOps Hub CUD Simulation

The FinOps Hub now offers an enhanced simulation feature for consumption-based CUDs (Committed Use Discounts). You can use a CUD recommendation as a starting point to simulate different usage scenarios and adjust the recommendation to maximize savings when purchasing a CUD. Learn more.

November 2023

BigQuery cross-region replication

Google BigQuery now supports cross-region replication, so you can replicate your data across different regions for disaster recovery, compliance and performance reasons. Cross-region replication is enabled by default for all new tables. However, the feature can also be enabled for existing tables. Read more.

GenAI EMEA Data Storage Residency

Google has launched GenAI Data Storage Residency in Belgium, the Netherlands, France, Germany and the United Kingdom. This means that customers can now store their data in EMEA and still benefit from Google's GenAI service. GenAI processing will continue to take place in the US, but with fully encrypted data. This solution will help many customers with their compliance requirements. Read more.

Updates August 2023

From August 29 to 31, Google announced no fewer than 161 updates to Google Cloud Platform and beyond at CloudNext. We have briefly compiled our top 6 innovations:

Duet AI for Google Workspace out of preview

Google is now making its AI capabilities available to everyone via Google Workspace as well. As always, this can of course be tried out in a 14-day trial. Duet AI will be supported in many products in the future. In Google Slides, for example, images can be generated with a simple click, which can be used directly in Slides and even customized. In Google Sheets, there is the "help me organize" function: This lets you generate entire templates to get your task done even faster. There is a lot more beyond that. Below are the two most relevant links to learn more: here and here.

Vertex AI Search and Conversation out of preview

First, the two products are out of Preview and generally available, and second, Google has renamed the two products. Until recently, they were known as Enterprise Search on Generative AI App Builder and Conversational AI on Generative AI App Builder. Of course, there are new features for both products as a result. There is now a "Multi-turn Search", which means that the search can be continued in the future if the result is not yet satisfactory. There is no need to start a new search, but it can be specified in a conversation with the search function. Another feature is the Vertex AI extensions, which can connect the AI directly to existing systems such as Salesforce, Confluence and JIRA. Data can be used much more easily this way. Read more.

New models in the Vertex AI Model Garden

Google has made new models and tools available for the Vertex AI Model Garden. These include Llama 2 and Code Llama, which are primarily developed by Meta. These are now available for anyone to use in the Google Cloud as well. In addition, Google has improved some of its own models, PaLM got a larger context window. The new 32,000 tokens can be used to evaluate larger documents. The model around images, Imagen, got a watermark function, so images generated with Imagen can be identified at any time. Read more.

Duet AI in preview for Looker, BigQuery and Dataplex

Duet AI is also now available in Google Cloud's Data Analytics products. First in preview, the AI launches in Looker, BigQuery and Dataplex products. In Looker, there will be a feature where you can "talk" to Looker using a chat window. Google refers to this as a virtual data analyst to whom you can ask questions about data and analysis. Google has now integrated BigQuery more closely with Vertex AI. This makes it much easier to run Generative AI on big data directly in BigQuery from now on. Read more.

New tool for BigQuery

In BigQuery, the biggest new feature is BigQuery Studio. This is a new interface in the Google Cloud Console where analytics, data engineering and predictions will take place in the future. Read more.

Duet AI also in the Security Command Center

AI is also playing an increasingly important role in security, including at Google. In the future, Duet AI will also be used in the Security Command Center. There, it is supposed to reduce complexity and help users more quickly recognize what is going on in the environment and whether it needs to be reacted to. The AI can identify where problems are occurring and give the user a collected overview of all affected components directly. Read more.

All updates from CloudNext are available on the Google Blog.

Updates July 2023

Cloud Run Jobs Runtime Limit Raised

Cloud Run Jobs now support significantly longer running jobs. These can now run for up to 24 hours. Read more.

More regions for Log Analytics

Log Analytics in Cloud Logging now supports new regions, making it available in almost all regions. Users can update their buckets to use Log Analytics in Singapore, Montréal, London, Tel Aviv, and Mumbai. Read more.

Data Residency available for Cloud Deploy

Cloud Deploy is now represented among Google Cloud services that can be configured to meet data residency requirements. This means that data is only stored and processed in pre-selected regions. Read more.

June 2023

Dataform is out of the preview

Dataform has been out of preview since June and is now fully ready for production use. The tool enables the complete management of SQL pipelines in BigQuery. These can be developed and rolled out in Dataform. Version control is also included. This means that the lifecycle of such pipelines can be mapped completely in one tool. Read more.

Customers can book SCC themselves

Recently, customers have been able to book the premium version of the Security Command Center (SCC) themselves via the console. This can be done without commitment and with flexible pay-as-you-go pricing. Read more.

Updates May 2023

Premium pricing for SCC is becoming cheaper

Security Command Center (SCC) premium pricing will be 25% cheaper at the project level. This affects customers who use SCC to secure Compute Engine, GKE Autopilot, App Engine and Cloud SQL. Read more.

New filter options for Log Analytics in Cloud Logging

Google has introduced the ability for Log Analytics to restrict queries by date and time, providing the same functionality already known from Logs Explorer and Metrics Explorer. This also includes options such as presets, customizable start and end times, and relative times. Read more.

Cloud Workstations now available as GA

Cloud Workstations, which provide a fully managed and fully integrated development environment in the Google Cloud, are now GA. With these, developers can be productive very quickly and have an environment that is permanently stable and up to date. Read more.

Updates April 2023

New Generation of Confidential VMs

A new generation of confidential VMs called "AMD Secure Encrypted Virtualization-Secure Nested Paging (AMD SEV-SNP)" is now available in private preview. This provides a confidential computing technology available on general N2D machines, built on top of memory encryption and offering new hardware-based security protections such as strong memory integrity, encrypted register state and hardware-assisted remote attach. You can sign up to participate here.

BigQuery change data capture (CDC) in public preview

The BigQuery Change Data Capture (CDC) feature is now in public preview. This feature is fully managed and provides a method to process and apply UPSERT and DELETE operations directly to BigQuery tables in real-time. Read more.

Manage Log Analytics with Terraform

It is now possible to configure Log Analytics in cloud logging buckets and BigQuery datasets using Terraform modules. The following modules can be used for this purpose: google_logging_project_bucket_config & google_logging_linked_dataset.

Updates March 2023

Google Workspace more useful through AI use

Google has announced plans to make more use of AI in future versions of Google Workspace. This will allow emails to be replied to, summarized and prioritized. In Docs, you'll be able to brainstorm, proofread and rewrite. In Chart, workflows can be generated to get things done. And a lot more beyond that. A preview can be found here.

Spanner gets read replicas and zero-downtime moves

Google has added new features to Spanner. First, read-only replicas can now be configured. So clients in any region get best latencies. Secondly, there is now the possibility to adjust the configuration of Spanner without downtime. Google takes care that the database is continuously available. Furthermore Google has adjusted the prices for multi-regional configurations downwards. Read more.

Updates February 2023

Security Command Center (SCC) Project-based and pay as you go

For those who like to ensure security in their Google Cloud environment, the Security Command Center is a good solution. Until now, this was accompanied by a relatively rigid pricing model and a booking for the entire cloud environment. Google now offers much more flexible and possibly also more cost-effective options. On the one hand, SCC can now also be booked for individual projects. And with a new pay-as-you-go payment option, there is more flexibility instead of long commitments. Read more.

Google Event: Data Cloud & AI Summit on 29 March

Google has announced a digital event for the cloud and AI space. It will showcase the latest developments and products for IT leads, developers and data professionals. Google will also share how other companies are dealing with challenges in these areas and what solution will yield the best results. Click here to register.

Confidential GKE nodes have reached GA status

For compute-optimized C2D VMs, there are Confidential Nodes in GKE. These have now reached GA status. This means that the use of Compute in the Google Cloud can be secured and protected a little further. The corresponding hardware offers encryption of the data in the RAM to additionally prevent data leaks. Read more.

Updates January 2023

Guide for the use of fast and cost-efficient Dataproc jobs

Google has published a Dataproc Guide on its blog. This explains how Dataproc, i.e. the Google Cloud Service to run various Apache tools, can most effectively be used. In particular, the authors address the optimization of existing clusters so that they generate less cost for the same result.

Guide for a reliable infrastructure design in the Google Cloud

This second guide from Google describes how to design and build infrastructure in the Google Cloud. The focus of this guide is a stable and reliable environment that can accommodate outages. It explains the differences of Google Cloud Datacenter regions and zones and how to make the most use of them in a design. 

General Availability of GPU Pods for GKE Autopilot 

Workloads that leverage the use of a GPU can now be used in GKE Autopilot. For example, this could be ML training, video encoding, or other jobs. Read more.

General Availability of Kubernetes v1.26 for GKE

The new Kubernetes version 1.26 is now available on GKE. This allows customers to deploy all the new features in the release and use the latest version. Google Cloud has set a goal of making a new Kubernetes OSS release available to customers within 30 days. Read more.

Updates November 2022

Popular trainings on Coursera free for a short time

Until the end of the year, some popular courses and trainings on Coursera are available to everyone for free. With this offer, both technical and non-technical are available to everyone. Certificates can also be taken afterwards. Read more.

Apigee Repeats Leader in 2022 Gartner Magic Quadrant for API Management

For the seventh year in a row, Apigee is a Leader for API Management. At a time when APIs are becoming increasingly important, Google continues to lead the market with Apigee. Through Apigee's innovations, enterprises can better manage their APIs, make them more secure, better integrated, and developers can better interact with them. Read more.

Updates October 2022

Guide for BigQuery and Log Analytics

Google created a guide for using Log Analytics with BigQuery. It compares the two solutions and shows the path from logs in BQ to Log Analytics with BQ. Read more.

General Availability of IAM Deny 

IAM Deny works like a guardrail for permissions. Using IAM Deny policies, access to specific Google Cloud resources can be restricted or denied depending on account properties, resource types and permission used. Administrators thus have another tool to increase security. Read more.

Cloud Spanner Lock and Transaction Insights

The new Cloud Spanner Lock and Transaction Insights features provide insights into lock conflicts and transactions with preconfigured dashboards. The dashboards make troubleshooting conflicts easier and enable troubleshooting of a slow database. Read more.

Updates September 2022

New product Cloud Logging - Log Analytics

Google Cloud presented a new product in preview that combines the benefits of Cloud Logging and Big Query. With Log Analytics, the logs available in Cloud Logging can be analyzed in Big Query. In doing so, the logs do not have to be re-imported, but can be used directly from Cloud Logging by Big Query. Read more.

Datastream News

PostgreSQL databases can now be used for Datastream. This feature is available in public preview. Read more.

Billing for data volumes in Datastream is now based on pricing tiers. The more data in total is added to the system, the lower the price per GB will be. Read more.

Pub/Sub Monitoring Dashboards

The Pub/Sub Monitoring Dashboards are now available in the GCP console. The dashboards, which provide helpful insights into the status of services, can now be viewed and edited in the Google Cloud Console. Read more.

New pricing model for Apigee

Previously, Apigee could only be used via completed subscription. Now the offering can also be obtained via pay-as-you-go. Google is thus offering a much more flexible model without a long-term commitment. The threshold for testing the offer is much lower than before. Read more.

Updates August 2022

2nd generation Cloud Functions are now generally available

The new generation of Cloud Functions offer a wider variety of workloads, more compute power, granular controls, and and all-around improved developer experience.

Learn more

Kubernetes and Skaffold configurations can now be viewed and compared in the GCP Console.

Learn more

New "Easy Mode" in Cloud Deploy

Cloud Deploy now offers an "Easy Mode." With this, the skaffold.yalm file can be automatically created from a Kubernetes manifest. The feature can be used via the Command Line and the generated file is optimal for learning and presenting GCP Cloud Deploy.

Learn more

 

Stop Cloud Deploy's Delivery Pipeline

Cloud Deploy's delivery pipeline can now be paused. If a problem occurs with a release, the pipeline can now be paused to ensure no further action happens.

Learn more

New flag in Cloud SQL prevents accidental deletion of instances

Cloud SQL now provides a way to prevent accidental deletion of an instance by setting another flag. The flag is set by default and if it is active, then it must first be deactivated in order to delete an instance of MySQL, Postgres or a SQL Server. This way, such an instance can no longer happen accidentally.

Learn more

Bigtable-BigQuery Federation is available as GA release

This allows queries to be made from BigQuery directly from Bigtable. No ETL is required for this.

Learn more

BigLake is available as GA release

This tool can be used to leverage data from multiple sources. BigLake can work with multicloud data, data stores, "data lakes" and warehouses.

Learn more

Updates July 2022

PostgreSQL compatible interface for Cloud Spanner

Postgres has become increasingly popular over the past few years. The database management system is used in many open source projects and custom software. With the new interface, the advantages of Cloud Spanner are now also accessible to developers with Postgres expertise. Google's SQL database offers fast, worldwide consistency, scales almost endlessly, offers 99.999% availability and requires no maintenance.

 

New Google tool to compare services from different cloud providers

Google's comparison tool provides a quick overview of generally available services in the cloud. It offers practical help for users who are familiar with Azure, for example, and are looking for a corresponding product in Google Cloud Platform, or for users who want to migrate to GCP with Azure or AWS Clouds.

 

Best practice for incorporating cloud security early

This article from Google describes the important role of considering security and privacy early in the planning of cloud projects and presents some best practices with Google products.

 

BigQuery BI Engine update: Preferred tables

In BigQuery BI Engine, “preferred tables” can now be defined, for example specific tables from which data is frequently needed or from which the data should be available more quickly. These are prioritized and query performance is improved. This saves resources and costs compared to caching all tables in BI Engine.

 

Current Consumption available in the GCP Console app

In the GCP Console App you can now also access the billing information of your current console consumption.