この日はコンピューティングと生成 AI 関連の話題が中心でした。Amazon EC2 では Graviton5 搭載の M9g/M9gd インスタンスが一般提供となり、U7i-8TB が欧州 (パリ) で利用可能に、EC2 Capacity Blocks for ML が AWS GovCloud (US) に拡大しました。生成 AI では Amazon Bedrock を中心とした知的文書処理パイプラインや、Agent-EvalKit による AI エージェント評価、Amazon Quick と Snowflake Cortex AI の MCP 連携などが紹介されました。運用面では Amazon Aurora の PostgreSQL 18 対応、CloudWatch Application Signals の強化、Elastic Beanstalk と CloudWatch Logs の統合、AWS Lake Formation の S3 直接アクセスなどが発表されました。日本語ブログでは AWS Summit Japan 2026 のブース紹介や GENIAC 第4期支援、Nitro System の形式検証などが取り上げられました。
EC2 新インスタンス: Graviton5 搭載 M9g/M9gd の一般提供、U7i-8TB のパリ提供、Capacity Blocks の GovCloud 拡大
生成 AI/文書処理: Bedrock を用いた IDP パイプライン、Agent-EvalKit、Quick と Snowflake Cortex AI 連携
データベース: Amazon Aurora PostgreSQL 18 対応
運用/監視: CloudWatch Application Signals 強化、Elastic Beanstalk の CloudWatch Logs 統合、MWAA Serverless の EventBridge 対応
セキュリティ/基盤: AWS Nitro Isolation Engine のハイパーバイザ形式検証
日本の取り組み: AWS Summit Japan 2026 ブース、GENIAC 第4期支援開始
AWS Lake Formation now enables you to read and write the underlying data files in Amazon S3 for tables registered in the AWS Glue Data Catalog. This provides you with a single set of permissions for both SQL queries and direct file access using your existing Lake Formation table grants.
With this launch, Lake Formation provides temporary, scoped credentials for registered S3 locations based on your table permissions. SELECT permissions grant read access, and SUPER permissions grant read and write access to the data at that location. This capability comes built-in with Amazon EMR 7.13 or later. As a result, you can access data files directly from your Spark jobs for tasks that require file level access such as model training, feature engineering, or debugging data quality issues.
You can also integrate your Apache Spark or Trino applications using APIs or through an open source plugin provided by AWS. Additionally, all access is logged in AWS CloudTrail to provide a unified audit trail across SQL and file-based operations on your tables.
This feature is available at no additional charge in all AWS Regions where AWS Lake Formation is available. To learn more, see Lake Formation documentation, EMR documentation, API reference, and open source plug-in.
AWS Elastic Beanstalk now provides a CloudWatch Logs integration directly in the environment Logs tab of the Elastic Beanstalk console. Previously, customers had to navigate to the CloudWatch console to find the relevant log groups and log streams for their environments. With this launch, customers can view CloudWatch log events without leaving the Elastic Beanstalk console.
The Logs tab displays log groups that an environment streams logs to, as well as log groups matching the aws/elasticbeanstalk/<env-name>/* prefix. Customers can select a log group to view its log streams, with the most recently active stream selected by default. A log stream dropdown allows switching between streams and filtering results. For deeper analysis, a View in CloudWatch dropdown provides direct links to the log group, log stream, or CloudWatch Logs Insights in the CloudWatch console.
This feature is available across all Elastic Beanstalk platform branches in all AWS Commercial Regions and AWS GovCloud (US) Regions where Elastic Beanstalk is available. For a complete list of supported Regions, see AWS Regions.
For more information about using Elastic Beanstalk with Amazon CloudWatch, see the AWS Elastic Beanstalk developer guide. To learn more, visit the AWS Elastic Beanstalk product page.
Amazon Managed Workflows for Apache Airflow (MWAA) Serverless now supports workflow and task state change events to Amazon EventBridge, enabling data engineering and platform teams to build event-driven automation for their Apache Airflow workflows.
Previously, monitoring workflow execution required custom polling logic or manual observation. With this launch, MWAA Serverless can emit events when workflows transition between states, including started, running, succeeded, or failed, and when individual tasks change state, such as scheduled, succeeded, failed, or up for retry. With this feature, you can further automate your existing workflows - for example, using EventBridge notifications to trigger alerts when a production workflow fails, automatically restart dependent pipelines when an upstream workflow succeeds, or log state transitions to Amazon S3 for compliance and auditing.
This feature is available in all AWS Regions where Amazon MWAA Serverless is available. For the complete list of supported Regions, see Regions in the Amazon MWAA Serverless User Guide. For pricing details, see Amazon EventBridge pricing.
To learn more, see Monitoring Amazon MWAA Serverless in the Amazon MWAA Serverless User Guide and Amazon MWAA Serverless events in the Amazon EventBridge Events Reference.
Amazon Managed Workflows for Apache Airflow (MWAA) Serverless now supports workflow and task state change events to Amazon EventBridge, enabling data engineering and platform teams to build event-driven automation for their Apache Airflow workflows.
Previously, monitoring workflow execution required custom polling logic or manual observation. With this launch, MWAA Serverless can emit events when workflows transition between states, including started, running, succeeded, or failed, and when individual tasks change state, such as scheduled, succeeded, failed, or up for retry. With this feature, you can further automate your existing workflows - for example, using EventBridge notifications to trigger alerts when a production workflow fails, automatically restart dependent pipelines when an upstream workflow succeeds, or log state transitions to Amazon S3 for compliance and auditing.
This feature is available in all AWS Regions where Amazon MWAA Serverless is available. For the complete list of supported Regions, see Regions in the Amazon MWAA Serverless User Guide. For pricing details, see Amazon EventBridge pricing. To learn more, see Monitoring Amazon MWAA Serverless in the Amazon MWAA Serverless User Guide and Amazon MWAA Serverless events in the Amazon EventBridge Events Reference.
Amazon Aurora PostgreSQL-Compatible Edition now supports PostgreSQL major version 18, starting with version 18.3. This release brings community improvements to query performance and database management, and introduces support for pg_roaringbitmap, a new extension that performs fast, memory-efficient set operations on large collections of integers. This enables use cases such as audience segmentation, tag-based filtering, and permission checks directly in the database without application-layer processing.
PostgreSQL 18 introduces B-tree skip scans, which improve query performance, and reduce index storage and maintenance overhead. Major version upgrades now retain optimizer statistics, ensuring consistent query performance immediately after upgrading without waiting for statistics to be regenerated. Logical replication can now stream large transactions in parallel, reducing replication lag and keeping downstream systems more current. Please refer to the Amazon Aurora PostgreSQL release notes for details.
You can upgrade your database using several options including RDS Blue/Green deployments, upgrade in-place, or restoring a snapshot. Learn more about upgrading your database instances in the Amazon Aurora User Guide. Aurora PostgreSQL 18.3 is available in all commercial AWS Regions and AWS GovCloud (US) Regions.
Amazon Aurora is designed for unparalleled high performance and availability at global scale with full PostgreSQL and MySQL compatibility. It provides built-in security, continuous backups, serverless compute, up to 15 read replicas, automated multi-Region replication, and integrations with other AWS services. To get started with Amazon Aurora, take a look at our getting started page.
Amazon CloudWatch Application Signals introduces service health ranking on the application map and new infrastructure, logs, and traces tabs on the service overview page. These capabilities let operators triage unhealthy services and inspect the underlying compute environment, log snippets, and trace details in one place, making it easier to find root causes without switching tools.
Customers use Application Signals to monitor the health of distributed applications, but identifying why a service was unhealthy often required leaving CloudWatch to correlate infrastructure data across separate tools. The application map now ranks services by health and shows runtime indicators on service nodes for Amazon EKS, Amazon ECS, AWS Lambda, and Amazon EC2, along with a new infrastructure tab that surfaces the compute and runtime environment, its components, and curated default metrics with deep links to the relevant monitoring tools. In addition, the service overview page provides the infrastructure, logs, and traces tab, helping operators spot issues in context of their application. With health-ranked services on the application map and new infrastructure, logs, and traces tabs, operators can instantly identify their most degraded services and drill into the compute environment, error-producing log snippets, and slow or failing transactions — all without leaving Application Signals. These capabilities span workloads running on Amazon EKS, Amazon ECS, AWS Lambda, and Amazon EC2, giving teams a single pane to move from symptom to root cause in minutes instead of hours.
These capabilities are available in all AWS Regions where Amazon CloudWatch Application Signals is supported. To learn more about this feature, see the Amazon CloudWatch Application Signals documentation . For pricing details, see the Amazon CloudWatch pricing page
Today, AWS is expanding support for Amazon Elastic Kubernetes Service (EKS) local clusters on AWS Outposts to first-generation and second-generation AWS Outposts racks running Amazon EC2 instances that boot from Amazon EC2 instance store. AWS Outposts offers static stability for Amazon EC2 instances backed by EC2 instance store, and AWS is now extending that benefit to Amazon EKS local clusters customers. With local clusters, the entire Kubernetes control plane runs on AWS Outposts, supporting advanced data residency requirements and mitigating the risk of impact from temporary network disconnects to the cloud.
Amazon EKS local clusters on AWS Outposts backed by Amazon EC2 instance store use an updated architecture that brings greater operational and feature-level parity with Amazon EKS clusters in the cloud. The Kubernetes control plane on your Outpost is managed by Amazon EKS in a service-owned account, so you don’t need to manage etcd backups or logging agents on control plane instances. New Kubernetes versions and Amazon EKS platform versions are made available for local clusters as they’re released for Amazon EKS in the cloud. Local clusters deployed with the updated architecture support Amazon EKS add-ons, IAM Roles for Service Accounts, EKS Pod Identity, OIDC authentication, access entries, and Bottlerocket worker nodes (in addition to Amazon Linux 2023).
The updated architecture and new capabilities are generally available on AWS Outposts racks backed by Amazon EC2 instance store in all commercial AWS Regions that support AWS Outposts racks. AWS Outposts that boot Amazon EC2 instances from Amazon EBS will continue to use the original local clusters architecture. For more information, see local clusters in the Amazon EKS user guide.
Amazon Quick now integrates with Snowflake Cortex AI through the Model Context Protocol (MCP), enabling teams to query their Snowflake data and documents using natural language, and automate multi-step workflows directly within their Quick workspace. After setting up the connection using Snowflake's managed MCP server with OAuth authentication, you can ask questions across structured data through Cortex Analyst and retrieve insights from unstructured documents through Cortex Search.
With this integration, you can build Flows in Quick that orchestrate Snowflake Cortex Agents to execute repeatable, governed workflows with consistent structured output. This is ideal for any multi-step process that spans structured data and unstructured documents. The same MCP connection is also accessible from Quick Chat and other Quick features. For example, users can ask ad-hoc follow-up questions or explore their Snowflake data conversationally alongside their automated flows. Quick intelligently routes relevant prompts to Snowflake Cortex AI and returns contextualized answers alongside enterprise knowledge stored in Quick Spaces, giving teams both the rigor of a structured process and the flexibility of a conversational interface.
The Snowflake Cortex AI integration with Amazon Quick is available in all AWS Regions where Amazon Quick is available.
Visit the Amazon Quick website to learn more and start your Quick free trial. To learn more about the Snowflake Cortex AI integration, read the blog. To learn more about Quick integrations, visit the integrations page.
Amazon EC2 High Memory U7i-8TB instances (u7i-8tb.112xlarge) are now available in AWS Europe (Paris) region. U7i instances are part of the AWS 7th generation and are powered by custom fourth-generation Intel Xeon Scalable processors (Sapphire Rapids). U7i-8TB instances offer 8 TiB of DDR5 memory, enabling customers to scale transaction processing throughput in a fast-growing data environment. U7i instances offer up to 45% better price performance over existing U-1 instances.
U7i-8TB instances deliver 448 vCPUs and support up to 100 Gbps of Amazon EBS bandwidth for faster data loading and backups, 100 Gbps of network bandwidth, and ENA Express. U7i instances are ideal for customers running mission-critical in-memory databases like SAP HANA, Oracle, and SQL Server.
To learn more about U7i instances, visit the High Memory instances page.
Amazon EC2 Capacity Blocks for ML is now available in AWS GovCloud (US-West) and AWS GovCloud (US-East), enabling government and regulated-industry customers to reserve GPU capacity for machine learning workloads.
EC2 Capacity Blocks for ML allows you to reserve GPU instances in advance for a defined duration, giving you assured access to accelerated compute for short-duration pre-training, fine-tuning, rapid prototyping, and inference demand surges. Capacity Blocks deliver low-latency, high-throughput connectivity through colocation in Amazon EC2 UltraClusters.
You can reserve capacity up to eight weeks in advance for durations up to 6 months, in cluster sizes of one to 64 instances. Capacity Blocks can also be shared across multiple accounts using AWS Resource Access Manager (RAM), helping organizations coordinate ML infrastructure investments and keep reserved capacity in continuous use across workloads.
In AWS GovCloud (US), EC2 Capacity Blocks for ML is available on P6-B200 instances in AWS GovCloud (US-West), and P6-B200 and P6-B300 instances in AWS GovCloud (US-East). To get started, visit the EC2 Capacity Blocks documentation.
AWS Graviton プロセッサは世代を重ねるごとに着実に進化を遂げ、イテレーションを経るごとに、コンピュ […]
国内最大規模の学習型 IT カンファレンスである AWS Summit Japan が、2026 年 6 月 […]
こんにちは。Amazon Web Services Japan のソリューションアーキテクト、田中 里絵 です […]
AWS DMS、Amazon S3、AWS Glue、AWS Step Functions を使って、Oracle から Amazon Aurora DSQL へデータを移行する手順を解説します。自動化されたコスト効率の高い移行パイプラインを構築し、Amazon Aurora DSQL のサーバーレスアーキテクチャ特有の課題に対応します。
2026年6月4日、経済産業省と国立研究開発法人新エネルギー・産業技術総合開発機構 (NEDO) が実施する […]
Blueprint instruction optimization is a BDA feature that automatically refines your extraction instructions to address this challenge directly. You provide three to ten example documents with expected values, and BDA refines your blueprint instructions to improve accuracy in minutes, not weeks. No separate model fine-tuning is required. By the end of this post, you can optimize your blueprints to improve accuracy, run the optimization workflow through the Amazon Bedrock console or the API, and apply best practices for selecting examples and ground truth.
Today, we’re excited to announce two new capabilities that make Quick Sight dashboards even more expressive and business-aligned: sparklines and custom sort for controls. In this post, we walk through both features, what they are, when to use them, and how to configure them, with real-world scenarios that bring them together in a practical, decision-ready dashboard.
Agent-EvalKit is an open-source toolkit (Apache 2.0) that makes this evaluation infrastructure available by integrating with AI coding assistants, including Claude Code, Kiro CLI, and Kilo Code. This post walks through how Agent-EvalKit works across its six evaluation phases, using a travel research agent built with the Strands Agents SDK and Amazon Bedrock as a running example.
This post demonstrates an intelligent document processing pipeline that consists of both on-demand inference and batch inference options on Amazon Bedrock to enable the flexibility on the document processing time and cost.
AWS Professional Services (AWS ProServe) compressed engagement timelines from months to days, not by adding artificial intelligence (AI) tools to an existing process, but by fundamentally rebuilding how we deliver from the inside out. In this post, we share how AWS ProServe became a frontier team, the practices that enabled it, and what your engineering organization can take from our experience.
This post outlines the development of a cost-effective and scalable intelligent document processing pipeline on AWS, powered by Amazon Bedrock and its features. BDA is a managed service within Amazon Bedrock that automates the extraction of insights from documents. We demonstrate how BDA extracts and analyzes document content, while Strands Agent hosted on Amazon Bedrock AgentCore Runtime coordinate specialized processing tasks, and Amazon Bedrock Knowledge Base enable contextual understanding across multiple documents. By combining these capabilities within a unified architecture, organizations can transform their document processing workflows with minimal development effort.
This post shows how to build a custom meeting prep and follow-up assistant using Amazon Quick and Cisco Webex MCP servers. From a single prompt, the agent finds an upcoming Webex meeting, reviews prior meeting summaries and transcripts, and pulls related Vidcast highlights and transcript context. It then searches Webex message threads for unresolved follow-ups and creates a concise prep brief. After the meeting, the same assistant can summarize the discussion and identify action items. It can also find related Vidcast updates and draft a follow-up message for the right Webex space.
One of the most impactful ways to improve the ROI on your Amazon Elastic Compute Cloud (Amazon EC2) investment is rightsizing — when you match your instance types and sizes to the actual resource demands of your workloads. However, doing this manually across hundreds or thousands of instances is time-consuming and error-prone. AWS Compute Optimizer […]
Organizations in regulated industries or with strict information security requirements are increasingly looking to use generative AI. However, they often face a dilemma: how to utilize powerful models while keeping data strictly on-premises or within specific geographic boundaries. The solution lies in deploying self-managed Small Language Models (SLMs) on premises with AWS Outposts or in […]
Ali Saidi is a VP and Distinguished Engineer at AWS Millions of customers use the AWS Nitro System to protect their most sensitive workloads, and AWS is an industry leader in innovation to secure customer data. Helping our customers keep their data secure and confidential is our highest priority, and we continue to make investments […]