AWS News - 2026-08-27

AWS Glue 5.1 is now available in the AWS European Sovereign Cloud Region.

AWS Glue is a serverless, scalable data integration service that simplifies discovering, preparing, moving, and integrating data from multiple sources. AWS Glue 5.1 upgrades core engines to Apache Spark 3.5.6, Python 3.11, and Scala 2.12.18, bringing performance and security enhancements. This release also updates support for open table format libraries, including Apache Hudi 1.0.2, Apache Iceberg 1.10.0, and Delta Lake 3.3.2. Additionally, AWS Glue 5.1 introduces support for Apache Iceberg format version 3.0, adding default column values, deletion vectors for merge-on-read tables, multi-argument transforms, and row lineage tracking. This release extends AWS Lake Formation fine-grained access control to write operations - both DML and DDL - for Spark DataFrames and Spark SQL. Previously, this capability was limited to read operations only. AWS Glue 5.1 also adds full-table access control in Apache Spark for Apache Hudi and Delta Lake tables, providing more comprehensive security options for your data.

With this expansion, AWS Glue 5.1 is now available all AWS commercial and AWS GovCloud (US) Regions.

You can get started with AWS Glue 5.1 using AWS APIs, AWS CLI, AWS SDK, or AWS Glue Studio. To learn more, visit the AWS Glue product page and our documentation. 

Amazon Connect Customer now supports points-based scoring to evaluate human and AI agent performance, giving managers more flexibility in configuring evaluation scores. With points-based scoring, managers assign points to each criterion based on its importance to the business, and the score is the sum of the points earned. In contrast, with the existing percentage-based scoring, each question is assigned weights that need to add up to 100%. For example, a points-based evaluation might award 40 points for resolving the customer's problem, 20 for meeting compliance requirements, and 10 for a proper greeting – without the need for scores to add up to 100. Managers control how points are awarded: combining several compliance criteria into one multiple-choice question to shorten the form, excluding the call reason from scoring, or awarding bonus points for strong customer rapport. Managers decide what counts, how much it counts, and what earns extra credit, so the score reflects what the business values most.

This feature is available in all regions where Amazon Connect Customer is offered. To learn more, please visit our documentation and our webpage.

Mountpoint for Amazon S3 can now limit memory usage, either automatically based on the environment it runs in or with a limit that you define. This lets you run Mountpoint alongside memory-intensive applications, for example, in machine learning training or analytics workloads where applications share a memory budget.

Previously, Mountpoint’s memory usage could expand over time based on usage patterns, potentially causing performance or stability issues when competing with other memory-intensive applications. With this launch, you can define a memory target for Mountpoint to reserve memory for your applications. Alternatively, Mountpoint can automatically determine a safe default based on the environment it runs in. For example, when running in Amazon EKS, Mountpoint can automatically detect a container's assigned memory budget. Under memory pressure, Mountpoint slows down operations to stay within this budget, enabling you to run Mountpoint in containers with strict memory allocations.  

Mountpoint is available in all AWS Regions. To upgrade to the latest version, visit the Mountpoint GitHub repository. To learn more about Mountpoint, see the overview page and the configuration guide in the Mountpoint GitHub repository.

Amazon Connect Customer now enables managers to input unplanned shrinkage in agent schedules, providing a more accurate picture of staffing due to unscheduled agent absences such as late logins or unplanned sick leave. Managers can upload unplanned shrinkage assumptions directly into agent schedules and immediately see updated scheduling metrics such as scheduled headcount, net staffing, and projected service level — each adjusted for unplanned shrinkage. For example, if 10% of the agents scheduled to work at 8 AM next Monday are expected to be unavailable due to late logins, then projected service level falls from 90% to 85% with unplanned shrinkage. This enables workforce managers to proactively address gaps in staffing that may arise due to unplanned agent unavailability, thus improving scheduling accuracy and their ability to maintain target service levels.

This feature is available in all  AWS Regions  where Amazon Connect Customer agent scheduling is available. To learn more about Amazon Connect Customer agent scheduling, click  here .

Amazon Cognito now provides a new admin API operation to reset a user's time-based one-time Password (TOTP) multi-factor authentication (MFA) configuration. When users lose access to their TOTP device, administrators can remove the device association, allowing the user to enroll a new device on their next sign-in.

This removes the need to recreate accounts to recover locked-out users if they lose access to their TOTP device. Customers can maintain MFA enforcement while providing a recovery path.

This new capability is available in all AWS Regions where Amazon Cognito is available. To get started, access the AdminDeleteSoftwareToken API using the AWS CLI, SDKs, or APIs. See the developer guide for instructions.

Amazon Aurora DSQL now lets you add foreign key constraints to new and existing tables. Aurora DSQL is a serverless, distributed SQL database with PostgreSQL compatibility and active-active multi-Region availability.

You can now express your application's rules about how tables reference each other as FOREIGN KEY constraints in your Aurora DSQL cluster. Declare, for example, that a customer's primary address must reference an existing row in your address table. Aurora DSQL then refuses any write that would leave that customer pointing at an address that no longer exists. You choose what happens when a referenced row is deleted or updated, with the options NO ACTION, RESTRICT, CASCADE, SET NULL, and SET DEFAULT.

This feature is available in all AWS Regions where Aurora DSQL is available. To learn more, see working with foreign key constraints and the CREATE TABLE foreign key syntax in the Aurora DSQL User Guide.

7 月は AWS オブザーバビリティにとってアップデートの多い月となりました。すでに収集しているテレメトリをより実用的なものにし、その収集にかかる運用作業を AWS 側で引き受ける機能をリリースしました。ログ分析は、ログクエリから直接実行できるアラームと、取り込み時に行われるエンリッチメントによって、アクションに繋げやすくなりました。アプリケーションレベルのオブザーバビリティは、エラーとデプロイイベントを自動的に捕捉するようになりました。AI コーディングエージェントの可視化が、新しいオブザーバビリティのカテゴリとして登場しました。Prometheus メトリクスのスクレイピングのために自前で運用していたエージェントは、マネージドコレクションによって不要になりました。そしてこれらすべてを支える基盤として、OpenTelemetry と Prometheus が引き続き中核的な役割を果たしています。 この記事では、Amazon CloudWatch、Amazon Managed Service for Prometheus、Amazon Managed Grafana、AWS DevOps Agent の最新情報をご紹介します。

Amazon はオープンソース分析データベース DuckDB を開発するアムステルダム拠点の DuckLabs を買収する最終契約を締結しました。DuckDB は 1 テラバイト以下の日常的なクエリを圧倒的な速度で処理するアーキテクチャを持ち、データエンジニアリング、データサイエンス、分析、AI エージェントの分野で広く採用されています。オープンソースプロジェクトは引き続き独立した Foundation の下で MIT ライセンスで維持され、AWS は S3 や Redshift、Athena などのサービスと DuckDB の強みを組み合わせていきます。

AWS にいる間、私はいつも学生と一緒に働く機会を探していました。私は地域の大学で 50 回以上の講演を行って […]

アマゾン ウェブ サービス ジャパン合同会社（以下、AWS ジャパン）は 2026 年 7 月 17 日（金） […]

A single alert from one security service tells you something happened. Read that signal alongside activity from other services and your own business context, and you will know whether what happened is part of a multi-stage attack. Consider a short sequence. An identity calls GetCallerIdentity from a source address it hasn’t previously used. Within minutes, […]

If you found time for a bit of vacation this summer, you might be in catch-up mode. Here’s a list to help: all the expert blog posts, new service capabilities, code samples, and workshops, in case you missed it, from July 2026. AWS Security Blog post This month’s AWS Security Blog posts covered AI agent […]

Most teams give AI agents either full access or read-only, leaving value unused or risk unmanaged. This post describes graduated autonomy, an architectural pattern in which agents earn expanded permissions through sustained reliability and lose them when performance degrades, built on Amazon Bedrock AgentCore, Amazon DynamoDB, and AWS CodePipeline.

Gallup transformed 90 years of workplace science into Gallup AI, a generative AI assistant powered by Amazon Bedrock that delivers real-time, personalized coaching to leaders directly within the Gallup Access application.

Learn how Amazon Bedrock AgentCore agents in one account can generate answers from an Amazon Bedrock knowledge base backed by Amazon Redshift Serverless in another account, without copying source data. This post covers the architecture, security boundary, and two orchestration models: a code-based Strands agent and a declarative AgentCore harness.

Data preparation determines the ceiling of any supervised fine-tuning project. This first post in a two-part series covers the foundations of SFT data prep: quality checks, conversational (JSONL) formatting, reasoning and tool-calling schemas, and a representative train/evaluation split.

The advanced side of supervised fine-tuning data prep. This second post in a two-part series covers evaluating data readiness with learning curves, selecting high-value data subsets, augmenting data with synthetic and distilled examples, and mixing data sources to prevent catastrophic forgetting.

The SageMaker Python SDK v3 redesigns script mode with unified ModelTrainer and ModelBuilder classes. This post walks through two end-to-end examples, a scikit-learn Random Forest and a multi-GPU Stable Diffusion 3.5 LoRA fine-tune, showing how SourceCode syncs your local code into any container at runtime so you can iterate without rebuilding Docker images.

Learn how Natera built an automated voice agent on Amazon Bedrock AgentCore that lets patients book mobile phlebotomy appointments through natural conversation. The post covers the dual-WebSocket bridge, event-driven latency masking, and progressive-trust authentication behind 100% tool-calling accuracy and sub-7-second latency.

In this post, you will learn how GoDaddy migrated from their legacy business intelligence (BI) tool to Amazon Quick. This was a two-year transformation that delivered results across every dimension of the business: 15,000 hours saved annually, 50% reduction in dashboard count, rendering times cut to under 5 seconds, and AI-powered self-service analytics now accessible to every employee.

Amazon Bedrock AgentCore Evaluations decouples agent evaluation from the framework you build on. As long as your agent emits OpenTelemetry telemetry, the service can score it, whether you use LangGraph, LlamaIndex, the OpenAI Agents SDK, Google ADK, the Claude Agent SDK, or Strands Agents. This post explains how the framework-agnostic contract works.