11/17/2026, 12:00:00 AM ~ 11/18/2026, 12:00:00 AM (UTC)

Recent Announcements

Amazon EMR 8.2 is now available

Amazon EMR 8.2 is now available with new features across Amazon EMR on EC2, Amazon EMR on EKS, and Amazon EMR Serverless. This release upgrades Apache Spark to 4.1.3, Apache Iceberg to 1.11.0, and Delta Lake to 4.2.0, and it is designated a Long Term Support (LTS) release. This release makes it easier to run secure, governed analytics at scale — you can now write to the same Delta Lake table on Amazon S3 from multiple clusters concurrently, refresh Iceberg materialized views incrementally, and propagate end-user identity from Spark jobs and interactive sessions all the way to AWS Lake Formation and Amazon S3 so access is authorized and audited as the original user.\n Amazon EMR 8.2 expands governance and performance across all three deployment options. New features include concurrent Delta Lake writes using Amazon S3 conditional writes (no DynamoDB log store required), and Lake Formation fine-grained access control (FGAC) now covers materialized views, governed user-defined functions, more Iceberg operations, and a broader set of Spark SQL features. The release also adds OpenLineage-based lineage capture to the AWS Glue Data Catalog and trusted identity propagation for steps and sessions. With EMR 8.2, data engineers can now precompute each pipeline stage as an Iceberg materialized view in SQL. Each materialized view refreshes only the changes instead of recomputing the entire source table across projections, inner joins, filters and SUM, COUNT aggregations. Specific to EMR on EC2, this release adds Spark Connect support on multi-primary clusters, along with Secret Agent improvements that cut user-provisioning latency by up to 97% under high concurrency. On EMR Serverless, new scaling profiles for Spark (performance-optimized, balanced, or cost-optimized) let you tune how quickly executors scale, and applications can now run Lake Formation-governed Glue UDFs. On EMR on EKS, the Spark Operator is upgraded to version 2 (based on open-source Spark Operator 2.5.0).

Amazon EMR 8.2 is available in all AWS Regions where Amazon EMR is available. To learn more, see the Amazon EMR 8.2 Release Guide for EMR on EC2, the EMR Serverless release notes, and the EMR on EKS release notes, or get started by creating a cluster from the Amazon EMR console.

Introducing the Apache Spark Optimization Agent for Amazon EMR Serverless

Amazon EMR Serverless now offers the Apache Spark optimization agent, which analyzes completed Spark job runs and delivers cost and performance optimization recommendations in seconds. You can select a cost or performance goal, and the agent returns specific configuration changes with the telemetry evidence and plain-language reasoning behind each suggestion. The agent can analyze a single run or compare it against a known-good baseline to pinpoint what changed. You access it from the EMR Serverless console or through the SageMaker Unified Studio MCP Server.\n The agent takes the job’s configuration, logs, and Spark telemetry together with your chosen intent and returns recommendations computed by a deterministic recommendation engine. The engine reads all job artifacts together to identify what is driving runtime and cost: too little parallelism, a skewed join key, memory pressure forcing spill, idle executors, or capacity provisioned beyond what the job uses. The configurations the engine targets are the settings most commonly tuned across millions of Spark jobs running on AWS. When cost and performance trade off, the recommendation improves the metric you asked for while keeping degradation of the other within defined limits. The engine also evaluates a best practices checklist covering platform-level improvements such as upgrading to a newer Spark version or using Graviton processors. Each recommendation is labeled by the surface it modifies: infrastructure, Spark configuration, code, or data layout. Data engineers can also use the agent during development to tune configurations before promoting to production. All access is authenticated with AWS IAM and logged in AWS CloudTrail.

The Apache Spark Optimization Agent is available at no additional charge in all Regions where EMR Serverless is available, except GovCloud (US) and China Regions, for EMR release versions 7.6 and later. To get started, see Apache Spark Optimization Agent for Amazon EMR Serverless.

AWS Transform now accelerates migration deals for Migration Acceleration Program (MAP) benefits

AWS Partners can now expedite Migration Acceleration Program (MAP) deals, from qualification to incentive payout, with an end-to-end agentic workflow in AWS Transform. Partners can manage the full MAP lifecycle without leaving AWS Transform, including deal qualification, funding approvals, agreement execution, migration planning and execution, resource tagging, and incentive disbursement.\nPreviously, Partners had to switch between separate systems for opportunity discovery, assessments, deal qualification, funding approvals, agreement execution, resource tagging, and incentive disbursement. This fragmentation, along with manual handoffs between steps, led to longer qualification cycles. Now, with the agentic MAP experience in AWS Transform, Partners connect their AWS Partner Central account and invite customers into the same engagement. Partners start by scoping customer environments with the built-in agentic discovery tool, or by importing data from RVTools or AWS-validated ISV tools listed in AWS Marketplace. AWS Transform then combines this discovery data with opportunity details from APN Customer Engagements (ACE) to automatically qualify Partner-initiated opportunities in a guided flow, so Partners get eligibility decisions sooner and close deals faster. As the migration progresses, MAP resource tags are applied automatically, and Partner cash and customer credits are disbursed based on actual usage data. Throughout, customers get shared, real-time visibility into migration projects alongside their Partner. AWS Transform with integrated MAP is now available in all AWS Regions where AWS Transform is supported. To learn more, visit the AWS Migration Acceleration Program (MAP) page. Partners can get started with the AWS Partner Central user guide (login required) or by contacting their Partner Development Manager. Customers can contact their AWS account team or AWS Partner.

Amazon ECR now supports global authentication tokens for private registries

Amazon Elastic Container Registry (Amazon ECR) now supports global authentication tokens for private registries. Previously, you needed to request a separate authentication token in each AWS Region where you push or pull container images. Now, you can request a single token from the AWS Sign-In CreateBearerToken API and use it across all ECR private registries in the same AWS partition.\n Teams that run proxy services to route container pulls across Regions, a common pattern for optimizing latency or availability, previously had to request, refresh, and coordinate a separate token for every Region behind the proxy. With global tokens, the proxy manages one token for all Regions in the AWS partition. Global tokens are valid for 12 hours by default, but can be configured to be as low as 5 minutes. These tokens work with the Docker CLI, and the Amazon ECR Docker Credential Helper. Existing regional authentication tokens continue to work as before.

Global authentication tokens are available in all commercial AWS Regions where Amazon ECR is supported. To learn more, see Private registry authentication in the Amazon ECR User Guide.

AWS Security Hub now infers resource criticality from installed software and usage

Today, AWS Security Hub adds installed software and resource usage as signals for resource criticality, which indicates how important a resource is to your business. Security Hub uses criticality to rank exposure findings, so that central security teams can prioritize the exposures that affect their most important resources.\n Security Hub already infers criticality from resource tags and names, but inconsistently tagged resources often have an Unknown criticality, which makes their exposures harder to prioritize. With installed software and resource usage as additional signals, more of your resources receive a meaningful criticality with no tagging or configuration required. For example, an EC2 instance that runs a database engine and serves steady traffic receives a higher criticality than an instance that runs only monitoring agents and sits idle.

These signals are included with the Security Hub Essentials plan at no additional cost and require no new enablement. To learn more, see Resources in Security Hub in the AWS Security Hub User Guide and the AWS Security Hub product page. For a list of AWS Regions where Security Hub is available, see the AWS Region table.

Amazon Bedrock AgentCore Memory now supports short-term memory retention of up to 10 years

Amazon Bedrock AgentCore Memory now lets you keep short-term memory events for up to 10 years, up from one year. Developers can now use short-term memory as a long-lived record of agent conversations and interactions. This helps when that history needs to stay available for a long time, such as for compliance and audit, for users who come back after months away, or for agents that need to look back at past sessions.\n You set retention on the memory resource, and it’s applied to each event when the event is created, so existing events keep their original retention. Standard short-term memory storage charges apply for as long as you keep the data.

This feature is available in all AWS Regions where AgentCore Memory is available. To get started, see the AgentCore Memory documentation.

Amazon API Gateway REST APIs now support JWT authorizers

Amazon API Gateway now offers a native JWT authorizer for REST APIs, a new authorizer type that validates JSON Web Tokens directly in the gateway. Until now, validating tokens from an OpenID Connect or OAuth 2.0 identity provider other than Amazon Cognito required a Lambda authorizer. You now configure token validation instead of writing and operating code for it.\n You can define one or more issuers, the audiences accepted for each, the signing algorithms you permit, and any additional claims a token may present. API Gateway verifies each token’s signature, issuer, audience, expiry, and claims on every request. You can attach one authorizer to many methods and require different scopes on each. This fits machine-to-machine APIs, mobile and web backends, and multi-tenant APIs that accept tokens from several identity providers.

The JWT authorizer works on Regional, edge-optimized, and private REST APIs, and is available in all commercial AWS Regions and the AWS GovCloud (US) Regions. To get started, see Amazon API Gateway documentation and AWS blog post.

Amazon Bedrock AgentCore Memory now provides full session summaries

Amazon Bedrock AgentCore Memory now keeps a full session summary for every session. It’s a single record that gives a snapshot of what happened in the conversation. Summary chunks cover the conversation in more detail, while the full session summary captures the overall picture of where the session stands. AgentCore Memory updates it as new events arrive, so it always reflects the session so far. This works well for developers who want a quick recap rather than a play-by-play.\n Full session summaries come with the summary strategy and need no extra setup. Every memory that has a summary strategy gets them automatically, including existing sessions. Each summary is stored as a long-term memory record next to the session’s summary chunks. You retrieve it with the Memory operations you already use, and you can filter by record type to return only full session summaries.

This feature is available in all AWS Regions where AgentCore Memory is available. To get started, see the Summary strategy documentation.

Amazon SageMaker now supports AI data preparation capabilities to build high-quality datasets for model development

Building high-quality training datasets has traditionally required weeks of manual labeling and data cleaning. Amazon SageMaker now offers AI data preparation capabilities that help data scientists and ML engineers build high-quality training and evaluation datasets for both text-based model customization and traditional ML.\n Teams can generate synthetic examples grounded in their own documents, auto-label records at scale with confidence scores, incorporate expert feedback through human review, evaluate data quality, and transform data for their chosen training or evaluation technique. Synthetic data generation creates training, evaluation, and test data from a specification grounded in your documents. Auto-labeling labels data at scale with built-in task types including classification, named entity recognition, extractive question answering, ordinal rating, and summarization, plus custom tasks. Rows below your confidence threshold go to a human review queue in the Label Studio Community Edition in SageMaker AI Studio. Data quality evaluation finds duplicates and surfaces issues such as tone and insufficient diversity with both quality metrics and rules. Data transformation cleans and reshapes data for your training or evaluation technique, using operators such as dedupe, filter, and format conversion.

Teams can start with just a few examples, production traces, or existing data in Amazon S3, Amazon Athena, or AWS Glue Data Catalog. Developers access these capabilities directly through the SageMaker Python SDK API or in natural language through the latest version of the AWS AI-ML skill in the Agent Toolkit for AWS. Your source data, generated datasets, and labels stay in your own Amazon S3 buckets, controlled by an IAM role you manage.

These capabilities are available in the US East (N. Virginia) (us-east-1) and US West (Oregon) (us-west-2) regions.

To learn more, see the AI data preparation documentation.

Amazon DynamoDB caching is now available with Amazon ElastiCache Serverless for Valkey

Today, AWS announces the general availability of DynamoDB caching with Amazon ElastiCache Serverless for Valkey, enabling sub-millisecond read latency for frequently accessed DynamoDB data. Previously, if you required faster reads than DynamoDB’s single-digit millisecond latency, you had to manually manage cache infrastructure or keep a separate cache in sync. Now you can access DynamoDB table data from a write-through cache that updates as data changes and automatically scales to meet changing workload demands.\n DynamoDB caching requires no application logic changes and minimal configuration. You choose which DynamoDB tables and APIs to cache and how long to retain cached data. Your application’s IAM credentials are used to access DynamoDB, so existing access control policies are automatically applied.

DynamoDB caching is available in all commercial and AWS GovCloud (US) Regions. DynamoDB caching is offered at no additional cost, so you pay only for the ElastiCache Serverless and DynamoDB resources your applications use. To get started, see the ElastiCache Developer Guide.

Amazon Bedrock AgentCore now offers free monthly web search queries

Amazon Bedrock AgentCore now offers 5,000 free web search queries per month through the Web Search Tool in AgentCore gateway and the Web Search API. AgentCore helps developers build, deploy, and operate highly effective agents securely at scale. Developers can now build and test agents with web search capabilities at no cost for their first 5,000 queries each month.\n The free allowance applies automatically, with no setup required. It is shared across supported AWS Regions and by all accounts in an organization, and resets each month. Additional queries are billed at standard rates without interrupting your agents.

Free web search queries are available in all AWS Regions that support the Web Search Tool and Web Search API: US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Ireland), Asia Pacific (Sydney), and Asia Pacific (Tokyo).

To learn more, see AgentCore pricing, Web Search Tool in the AgentCore Developer Guide, and the AgentCore product page.

AWS Config announces managed conformance packs

AWS Config now offers managed conformance packs. AWS builds and maintains these packs for specific compliance and governance frameworks. When you deploy one, you choose a framework, and AWS Config evaluates your resources against it. AWS keeps the pack up to date as things change. When new AWS Config managed rules map to the framework’s controls, or when the framework itself is revised, AWS updates your deployed pack automatically. Because AWS maintains a consistent pack definition in every supported Region, you can deploy the same framework across Regions without building or maintaining Region-specific rule sets. With managed conformance packs you no longer have to track framework changes or update rules yourself.\n Previously, to evaluate resources against a framework such as PCI DSS or SOC 2, you started from a sample conformance pack template, customized it for your environment, and then updated it yourself whenever new rules became available or the framework changed. Managed conformance packs remove that work, so your coverage stays current without manual template changes. Each pack includes the managed rules available in the Region where you deploy it and adds new rules as they become available there, so you don’t need a separate template for each Region. You can set rule parameters to fit your environment. Compliance results and compliance scores appear in the AWS Config console alongside your other rules and conformance packs.

Managed conformance packs cover frameworks including PCI DSS v4, SOC 2, NIST SP 800-53 Rev. 5, ISO/IEC 27001:2022, and CIS Controls, along with AI governance frameworks such as the EU AI Act and NIST AI RMF 1.0. For the full list, see the AWS Config documentation page.

Managed conformance packs are available in all AWS commercial Regions except Middle East (UAE) and Middle East (Bahrain). For the full list of supported Regions, see AWS Capabilities by Region. Evaluations are billed at standard conformance pack evaluation rates. To get started, open the AWS Config console and choose Conformance packs. To learn more, see the AWS Config Developer Guide and the AWS Config pricing page.

Amazon Aurora DSQL now offers SQL statement statistics in Database Insights

Amazon Aurora DSQL now records statistics for each SQL statement your cluster runs and publishes them as Amazon CloudWatch OpenTelemetry (OTel) metrics. You can view them in Amazon CloudWatch Database Insights. For each statement, you see how often it runs, how long it takes, the rows it scans, returns, or changes, and the DPUs it uses.\n Statement statistics measure every successful statement rather than sampling activity. Use them to find the statements that take the most time, run most often, or cost the most. In Database Insights, choose a statement under Top SQL to see its normalized SQL text and per-call charts. You can also query the metrics with PromQL in CloudWatch. Turn on statement statistics for each cluster in the Aurora DSQL console or the API when you create or update the cluster.

Statement statistics are available in all AWS Regions where Aurora DSQL is available. Standard CloudWatch metrics and Logs charges apply, with no additional Aurora DSQL charge. To learn more, see monitoring SQL statements with statement statistics in the Aurora DSQL User Guide and CloudWatch pricing.

Amazon EVS launches support for r7i.metal instances, external principal storage, and an Agent Skill

Today, we are announcing that Amazon Elastic VMware Service (Amazon EVS) now supports the r7i.metal-24xl and r7i.metal-48xl Amazon Elastic Compute Cloud (EC2) bare metal instance types for VMware Cloud Foundation (VCF) 9.x deployments. EC2 r7i instances are memory optimized, with custom 4th Generation Intel Xeon Scalable (Sapphire Rapids) processors, and designed to be used with external storage. With these instances possessing no local NVMe disks, Amazon EVS now supports VCF external principal storage with solutions such as Amazon FSx for NetApp ONTAP. With this launch, you can use the Amazon EVS Deployment Orchestrator (EDO) within the Solutions for EVS GitHub repository to fully automate the end-to-end deployment and configuration of your VCF 9.x environment on EVS.\n With this release, you now have more options for running VMware-based Virtual Machines (VM) in your Amazon VPC - while keeping the same tools, skills, and operational processes, and full control of your architecture, configuration, and lifecycle of your environment. 

We are also launching a new Amazon EVS Agent Skill, which gives your AI Agents context to help you plan, deploy, and operate your Amazon EVS environments - from answering questions and recommending designs to generating blueprints and driving deployments with EDO. To get started, point the AI agent of your choice to the EVS Agent Skill file in the latest release of the EVS GitHub repository.  

This release is available in all regions where Amazon EVS is offered. To learn more about Amazon EVS, visit the product detail page and the user guide.

OpenAI GPT-6.2 Astra, and GPT-6.2 Sol are now generally available on Amazon Bedrock

Today, AWS announces the general availability of OpenAI GPT-6.2 Astra and GPT-6.2 Sol on Amazon Bedrock. GPT-6.2 Astra brings frontier-level reasoning and judgment to your most ambitious work, including complex business problems, demanding research, and advanced coding and agentic workflows. Astra also includes an Ultrafast mode for latency-sensitive applications that need fast responses. GPT-6.2 Sol is designed for recurring coding, agentic, and professional workflows with clear outcomes, where speed and cost per task matter.\n Use GPT-6.2 Astra for challenging work that calls for expert judgment and a finished deliverable, such as reconciling competing inputs, investigating complex codebases, reasoning across text and images, and creating professional artifacts. Select ultrafast mode for Astra for interactive coding assistants, real-time agents, and customer-facing applications where latency shapes the customer experience. Use GPT-6.2 Sol for recurring, high-volume work with verifiable outcomes, such as implementing and debugging software, analyzing documents and data, and automating multistep workflows. Established AWS controls help you secure workloads, govern access, and audit model invocation activity. 

You can get started in the Amazon Bedrock console or programmatically through supported Amazon Bedrock APIs. For information about supported AWS Regions, endpoints, APIs, features, inference profiles and pricing, see the Amazon Bedrock documentation.

Amazon WorkSpaces Advisor now helps you find issues before your users notice them (Public Preview)

When a WorkSpace is slow or won’t connect, the IT administrator trying to help has traditionally had to dig through Amazon CloudWatch metrics, logs, and configuration data across multiple tools to find out why. And in the meantime, the user on the other end can’t get their work done. In April this year, we launched Amazon WorkSpaces Advisor, an AI-powered agent that helps IT administrators diagnose and fix WorkSpaces performance, connectivity and provisioning issues and get users back to work faster. With Advisor, administrators can investigate a WorkSpace from the console and, in minutes, see the likely root cause and recommended actions they can take right away. Since then, we’ve continued to invest in Advisor so your team can spend less time chasing down issues one by one, and more time on the projects your business is counting on you to deliver.\n Today, we’re excited to announce the preview of proactive troubleshooting in WorkSpaces Advisor, which moves your team from reacting to individual tickets to staying ahead of issues across your whole environment. Advisor can identify issues across your WorkSpaces, explain the likely cause of each one and recommend a fix. You don’t need to start an investigation, wait for a user to report a problem, or set alarm thresholds first.  

If the same problem affects many WorkSpaces, Advisor will show it as a single issue with the likely cause and a recommended fix.  Your team fixes the root cause once instead of chasing the same symptom one WorkSpace at a time. For example, Advisor can tell you when a group of WorkSpaces became unhealthy at the same time, when WorkSpaces are running low on CPU or memory, or when a certificate your users need to sign in is about to expire. That gives you time to act before a small issue turns into an outage or a queue of support tickets. You’ll find these issues in the WorkSpaces console.

Proactive troubleshooting also extends Advisor to your users, which means fewer tickets reach your help desk and the ones that do come with better information. When a session is slow or won’t connect, Advisor will tell users whether the problem is with their device, their network, or their WorkSpace. It then walks them through the fixes you’ve approved, such as restarting their WorkSpace. When an issue does need IT, the user gets a clear explanation to pass along, so your team starts from a diagnosis instead of “my WorkSpace is slow.”  You stay in control because you decide which fixes users can make on their own, and you can turn this off for any directory.

Proactive troubleshooting is available in preview at no additional charge in AWS commercial Regions where Amazon WorkSpaces Personal is available. Troubleshooting for end users requires version 5.37 or later of the WorkSpaces client for Windows, with the new experience turned on. Troubleshooting for users requires the WorkSpaces client for Windows version 5.37 or later with new experience. To get started, see the Amazon WorkSpaces Advisor Administration Guide.

Amazon VPC, AWS Transit Gateway, and AWS Cloud WAN now offer a 99.999% availability SLA for cross-Region traffic

AWS now offers a 99.999% availability service level agreement (SLA) for cross-Region traffic over Amazon Virtual Private Cloud (Amazon VPC) cross-Region peering, AWS Transit Gateway, and AWS Cloud WAN, with a new capability called Multipath Resilience. Multipath Resilience makes it easier to improve the resilience of mission-critical applications that send traffic across Regions, such as payment processing, financial trading, telephony, and live media streaming.\n The AWS backbone is highly resilient and connects Regions over physically diverse fiber paths. When a path fails, such as during a fiber cut, the network’s automatic sub-second rerouting moves traffic to a healthy path. In the rare event that such mechanisms do not resolve the issue in a timely manner, such as when there’s a partial degradation, it might take minutes for connectivity to be reestablished. Some applications, such as payment processing, financial trading, telephony, and live media streaming, can’t tolerate single-digit seconds interruptions, let alone minutes of unavailability. Until now, customers running these applications relied on workarounds, including modifying their applications and operating systems to split a single connection into multiple connections, or replicating data into every Region to avoid cross-Region calls. These workarounds add complexity and don’t always work, because applications can’t control which network paths their traffic takes.

Now, when you turn on Multipath Resilience for a cross-Region VPC peering connection, cross-Region Transit Gateway peering attachment, or cross-Region traffic on Cloud WAN, the network sends copies of each packet across physically diverse paths. It delivers whichever copy reaches your resources in the destination Region first and discards the duplicates. If one path drops packets, a copy is already in flight on another path, so your applications stay connected without waiting for the network protocols to reroute. Multipath Resilience moves this work out of your applications and into the network. Your applications keep using a single connection, with no changes to application code or operating systems, and no need to duplicate data to keep calls within a single Region.

Multipath Resilience is now generally available in US East (N. Virginia), US East (Ohio), US West (N. California), US West (Oregon), and Europe (Ireland).

To learn more, visit the Multipath Resilience documentation. For SLA details, visit the Multipath Resilience SLA. For pricing, visit the Amazon VPC pricing page.

AWS announces general availability of AWS Interconnect - multicloud connectivity with Microsoft Azure

AWS announces the general availability of AWS Interconnect - multicloud with Microsoft Azure. \n Customers have been adopting multicloud strategies while migrating more applications to the cloud. They do so for many reasons including interoperability requirements, the freedom to choose technology that best suits their needs, and the ability to build and deploy applications on any environment with greater ease and speed. Previously, when interconnecting workloads across multiple cloud providers (CSPs), customers had to go the route of a ‘do-it-yourself’ multicloud approach, leading to complexities of building and managing global multi-layered networks at scale. AWS Interconnect - multicloud is the first purpose-built product of its kind and a new way of how clouds connect and talk to each other, allowing customers to quickly provision resilient, scalable private connections to other cloud providers.

In August, Microsoft Azure launched support for AWS Interconnect in public preview and became the latest CSP to adopt the open specification that powers the service. With today’s GA launch, AWS customers can now rely on the same consistent, simple experience to interconnect their workloads on Microsoft Azure, Google Cloud, and Oracle Cloud Infrastructure (OCI).

Interconnect - multicloud is available with Microsoft Azure in the us-east-1 (N. Virginia), ap-southeast-2 (Sydney), and eu-central-1 (Frankfurt) AWS Regions. You can create an Interconnect using the AWS Management Console, Command Line Interface (CLI), or API. For more information, see the AWS Interconnect - multicloud documentation.

Amazon WorkSpaces provides a new Windows client experience

Amazon WorkSpaces provides a new Windows client experience for Amazon WorkSpaces Personal. The new client experience includes an updated interface, a simplified connection flow, easier-to-find settings, and the new WorkSpaces Advisor proactive troubleshooting (Public Preview) that suggests actions you can take to resolve common issue\n WorkSpaces Advisor proactive troubleshooting (Public Preview) is an intelligent new capability that monitors key metrics from your device, network, and the WorkSpace itself, and provides you with actionable recommendations if there are issues with registration, sign-in, or the connection in general. The new client experience also includes an improved full screen experience, a welcome tour to help orient you, and for WorkSpaces that are admin-enabled, a self-service option to enable nested virtualization.

The new client experience is available at no additional cost in all AWS Regions where Amazon WorkSpaces Personal is available.

To learn more, visit the Amazon WorkSpaces Windows client documentation or download the Amazon WorkSpaces client.

Amazon MSK Standard Brokers now support Graviton4-based M8g instances

AWS Graviton4-based M8g instances are now generally available for Standard Brokers in Amazon Managed Streaming for Apache Kafka (Amazon MSK). Compared to Standard Brokers on M7g, M8g delivers up to 35% lower P99 produce latency and 30% faster catch-up for consumers that fall behind.\n Standard Brokers on M8g are available today across all AWS regions where MSK Standard Brokers and M8g instances are available. For pricing, refer to the Amazon MSK pricing page. To get started, use the AWS Management Console or API to create a new cluster with Standard Brokers on M8g or start an in-place migration of an existing cluster from Standard Brokers on M7g to M8g. To learn more, see the Amazon MSK Developer Guide.

Amazon MSK Express Brokers now support Graviton4-based M8g instances

AWS Graviton4-based M8g instances are now generally available for Express Brokers in Amazon Managed Streaming for Apache Kafka (Amazon MSK). Compared to Express Brokers on M7g, M8g delivers 15% lower compute costs, up to 45% lower P99 produce latency, and 2x faster catch-up for consumers that fall behind. Additionally, customers no longer pay the broker-side cost when moving data across Availability Zones between clients and brokers, reducing total cross-AZ data transfer costs by 50%. Together, lower data transfer and broker-hour charges result in up to 30% cost savings compared to Express Brokers on M7g.\n MSK Express Brokers are designed to deliver up to 3x more throughput per broker, scale up to 20x faster, and reduce recovery time by 90% compared to Standard Brokers running Apache Kafka. Express Brokers on M8g further improve on M7g with better economics and performance along with existing Express capabilities. Customers running Apache Kafka on external clusters, including on-premises, self-managed on AWS, or other cloud providers, can use Amazon MSK Replicator to seamlessly migrate to Express Brokers on M8g and immediately benefit from the improved price and performance. Existing Express customers can upgrade from M7g to M8g through in-place migration while client applications keep producing and consuming without interruption.

Express Brokers on M8g are available today across all AWS regions where MSK Express Brokers and M8g instances are available. For pricing, refer to the Amazon MSK pricing page. To get started, use the AWS Management Console or API to create a new cluster with Express Brokers on M8g, start an in-place migration of an existing cluster from Express Brokers on M7g to M8g, or set up a replicator from an external Kafka cluster with Amazon MSK Replicator. To learn more, see the Amazon MSK Developer Guide.

Amazon API Gateway REST APIs now support JWT authorizers

Amazon API Gateway now offers a native JWT authorizer for REST APIs, a new authorizer type that validates JSON Web Tokens directly in the gateway. Until now, validating tokens from an OpenID Connect or OAuth 2.0 identity provider other than Amazon Cognito required a Lambda authorizer. You now configure token validation instead of writing and operating code for it.\n You can define one or more issuers, the audiences accepted for each, the signing algorithms you permit, and any additional claims a token may present. API Gateway verifies each token’s signature, issuer, audience, expiry, and claims on every request. You can attach one authorizer to many methods and require different scopes on each. This fits machine-to-machine APIs, mobile and web backends, and multi-tenant APIs that accept tokens from several identity providers.

The JWT authorizer works on Regional, edge-optimized, and private REST APIs, and is available in all commercial AWS Regions and the AWS GovCloud (US) Regions. To get started, see Amazon API Gateway documentation and AWS blog post.

AWS Security Hub extends unified security management to Google Cloud Platform

Today, AWS announces that AWS Security Hub now monitors Google Cloud Platform (GCP) resources, extending risk analytics, cloud security posture management, vulnerability management, and security response management across both clouds. Many AWS customers running workloads in AWS and GCP have had to operate separate security tools for each environment, making it difficult to prioritize risks holistically or respond consistently. Security Hub now provides a single, unified experience to detect and respond to risks across your AWS and GCP environments in addition to existing support for monitoring Microsoft Azure environments.\n Security Hub automatically discovers GCP resources, including GCP Virtual Machines (VMs), GCP Google Artifact Registry (GAR) container images, GCP Cloud Run functions, and GCP identities, and evaluates them for misconfigurations, internet exposure, and software vulnerabilities. You receive posture checks against security standards including the CIS Benchmarks™ for Google Cloud Platform Foundations, unified resource inventory, risk and exposure analysis, and automated response through existing EventBridge integrations. AWS and GCP findings appear in the same prioritized view with the same finding formats and automation workflows, so security teams can operate from one console rather than switching between tools.

Security Hub includes an independent 30-day free trial to monitor GCP resources that begins once you create your integration with Google Cloud Platform. After the trial, you pay the same price for monitoring GCP resources and equivalent AWS resources. You can create an integration to GCP from all AWS Regions where Security Hub is available except Middle East (UAE), Middle East (Bahrain), Asia Pacific (Taipei), and Asia Pacific (New Zealand). You can also create integrations to Google Cloud Platform for AWS Security Hub CSPM for posture management checks and Amazon Inspector for vulnerability management independently from AWS Security Hub. To learn more, see AWS Security Hub Pricing and AWS Security Hub documentation.

AWS launches credit splitting and transfers

AWS now supports credit splitting and transfers, enabling customers to divide any AWS credit into multiple smaller credits and transfer them to other linked accounts within their payer account. Previously, AWS applied credits starting with the owning account, then shared any excess with other accounts based on their usage, making it difficult for customers with large credits to direct precise amounts to specific teams or business units. FinOps teams can now split credits by dollar amounts, percentages, or even splits, and transfer each portion to the team that earned it.\n Credit splitting and transfers eliminate hours of monthly manual chargeback work for FinOps teams and provide predictable credit budgets per team or business unit. Each split credit inherits the same expiration date and service eligibility as the original, and split credits retain sharing settings by default so unused balances flow to other eligible accounts. Accounts receive first-priority consumption for transferred credits.

Credit splitting and transfers compose naturally with existing credit features, including credit-level sharing and credit pausing. After a split, customers have full control over each resulting credit. For example, customers can split and share one credit across all intended accounts so multiple teams benefit, while restricting another split credit to a single account. They can also pause a credit earmarked for a future project, then resume it when needed.

You can split any credit type via the AWS Billing Console, CLI, or API. Partially consumed credits can be split, with the split applying to remaining balance only. Each credit family supports up to 100 splits, with a $100 minimum per split credit. Splits are permanent with no undo operation.

Credit splitting and transfers are available in all AWS commercial regions at no additional cost.

AWS IAM announces Access Troubleshooter to debug access denied errors

AWS Identity and Access Management (IAM) introduces Access Troubleshooter, a new capability that gives you visibility into the full authorization context behind access denied errors so you can resolve them without trial-and-error. Previously, developers relied on the Amazon Resource Name (ARN) of the denying policy to debug a denial, with no visibility into the authorization context or other policies evaluated in the request that could contribute to the denial. This made troubleshooting access issues time-consuming and inefficient.\nStarting today, access denied error messages for supported services and APIs will include a unique identifier that you can use to retrieve details of the authorization performed by the service. You gain visibility into what policies were evaluated for each request, the authorization context of those requests, and the results of the evaluation. This enables you to identify the root cause of authorization issues quickly, streamlining your access management workflows and reducing time spent debugging permission problems. To learn more about Access Troubleshooter and how it can help you resolve access denied errors efficiently, see the User Guide on troubleshooting access denied error messages.

AWS announces agentic account creation for agentic coding harnesses

AWS now enables agentic coding harnesses to open the new AWS sign-up experience for developers who don’t have an AWS account yet. Previously, developers working in agentic coding environments like Claude Code, Codex, and Cursor had to manually navigate to aws.amazon.com, sign up for an AWS account and configure credentials before deploying workloads, interrupting the developer’s experience. Now, developers can go from an idea to deploying a running application by the browser automatically opening, completing sign up and returning short-lived credentials to the harness.\nWhen developers sign up for AWS from their coding agent, they start with the new AWS experience. Their first project comes set up with a home Region, IAM role manager, and short-term credentials that refresh automatically, so the agent can start deploying right away. To try the new experience, install the AWS plugin or configure the AWS Agent Toolkit in your agentic coding harness. To learn more, see the AWS Agent Toolkit documentation or the AWS Account Documentation.

Amazon Route 53 Resolver DNS Firewall now supports Infoblox Threat Defense (Preview)

Amazon Web Services announces the preview of Infoblox Threat Defense on Amazon Route 53 Resolver DNS Firewall. Security administrators can now apply Infoblox threat intelligence directly in DNS Firewall rules to preemptively protect applications in their Amazon Virtual Private Clouds (VPCs) from cyber threats, by subscribing to the Infoblox product listing on AWS Marketplace from the Route 53 DNS Firewall console.\n With this launch, you can add Infoblox managed rules to your DNS Firewall rule groups to filter and block queries to malicious and high-risk domains at the DNS layer. Infoblox continuously updates its threat intelligence feeds to preemptively protect cloud workloads from phishing, malware, ransomware and other advanced attacks. You can apply these protections to DNS queries from your VPCs and to hybrid traffic forwarded through Route 53 Resolver endpoints, so security teams keep consistent DNS policies across AWS and on-premises environments and add a layer of protection on top of AWS managed domain lists.

Infoblox Threat Defense on Route 53 DNS Firewall is available in preview in all AWS commercial Regions where Route 53 Resolver DNS Firewall is available. DNS Firewall Advanced customers can add Infoblox rules to existing rule groups at no additional DNS Firewall charge, and the Infoblox subscription in AWS Marketplace is free during preview.

To get started, see the Route 53 DNS Firewall documentation. To view Route 53 pricing, visit the Route 53 pricing page. To learn more about the AWS Marketplace listing and pricing for Infoblox Threat Defense for Route 53 DNS Firewall, see here.

AWS announces consistent sharing controls for credits, Savings Plans, and Reserved Instances

AWS Billing and Cost Management now gives customers symmetric, granular control over how credits, Reserved Instances, and Savings Plans are shared across their organization. Organizations with grant funding, regulatory obligations, or multi-entity structures can now automate benefit allocation that previously required manual chargeback reconciliation. With these new capabilities, customers can set default sharing preferences that automatically apply to all cost-saving benefits, and override those defaults for individual benefits when business requirements demand it.\n Customers configure Credit Group Sharing through the Billing Preferences page to set a single default that every new credit inherits automatically, ending the cycle of per-credit manual configuration. For Reserved Instances and Savings Plans, customers configure sharing preferences at the individual commitment level through the Savings Plans or Reserved Instances inventory pages, whether purchased at a linked account or at the payer level. All sharing configurations use Cost Categories, the same grouping construct customers already use in Cost Explorer, AWS Budgets, and the Cost and Usage Report.

These capabilities are available starting November 17, 2026, in all AWS Regions through the AWS Billing and Cost Management Console and public APIs. There is no additional charge. To learn more, visit the for documentations for Credit Group Sharing and Savings Plan and Reserved Instances Level Sharing.

AWS Split Cost Allocation Data now supports Amazon S3

Starting today, you can allocate the storage and request costs of a shared Amazon S3 general purpose bucket to the prefixes in that bucket in the AWS Cost and Usage Report (CUR). Organizations often store data for many teams, applications, or tenants in one bucket under different prefixes, such as marketing/ and finance/. Until now, attributing that bucket’s bill to each prefix required custom pipelines built on S3 Inventory or Storage Lens reports and your own pricing logic.\n With Split Cost Allocation Data for Amazon S3, Amazon S3 Storage Lens publishes prefix-level storage and request metrics to AWS Billing and Cost Management, which allocates the bucket’s storage and Tier 1 and Tier 2 request costs to the deepest matching prefix in the range of prefix levels you choose. Each prefix line item carries its full prefix path as AWS-generated cost allocation tags, so you can total cost by top-level prefix, team, or application, and reconcile it back to the bucket. You can use this data for chargeback, showback, and to see which prefixes drive your S3 spend.

To get started, enable AWS Billing and Cost Management publishing in an S3 Storage Lens advanced tier dashboard configuration, then enable split cost allocation data for Amazon S3 in the AWS Billing and Cost Management console. Prefix cost data appears in your CUR after the next daily Storage Lens metrics are published. Split Cost Allocation Data for Amazon S3 is available in all AWS commercial Regions where S3 Storage Lens advanced metrics are available, at no additional charge beyond S3 Storage Lens advanced tier pricing. To learn more, visit the AWS Cost and Usage Reports product page, Understanding split cost allocation data, and Estimating Amazon S3 costs by prefix with S3 Storage Lens.

Amazon SageMaker HyperPod now supports cluster operations for tracking scaling and on-demand Deep Health Check requests

Amazon SageMaker HyperPod introduces cluster operations, giving you visibility into every scaling and on-demand deep health check request from submission to completion. When a request runs into errors, the SageMaker AI console shows an alert that links directly to that request’s error events.\n With continuous provisioning, HyperPod brings capacity online in the background as it becomes available, retries failed nodes automatically, and lets you run multiple scaling requests at the same time. Cluster operations show you when each request finishes and which request caused the activity on your cluster. When a scale-up falls short of its target, you can go directly to the failure reason.

Each scaling or on-demand deep health check request creates a cluster operation that tracks the request from submission to completion, including whether it is still in progress and what it changed on each instance group. Events that HyperPod records while working on the request are linked to its operation. For example, if a scale-up of your training instance group doesn’t reach its target, you can open the operation and see only the events for that request, without sorting through unrelated cluster activity. If you or an autoscaler such as Karpenter send several requests to the same instance group, for example during a burst of inference traffic, each operation shows whether it completed or was replaced by a newer request. That way you can tell which request actually changed your cluster. After you run an on-demand deep health check before a large training job, you can follow its operation to confirm when it finished and which instances it affected. In the SageMaker AI console, operations appear on the cluster’s Events tab. To build the same visibility into your own dashboards, alerting, or CI/CD pipelines, use the new ListClusterOperations and DescribeClusterOperation APIs, and filter ListClusterEvents by operation.

Cluster operations are created automatically for every request, with no changes to your existing workflows.

Cluster operations are available at no additional cost for SageMaker HyperPod clusters that use continuous provisioning, in all AWS Regions where SageMaker HyperPod is available. To learn more, please check the documentation.

Amazon Route 53 Resolver DNS Firewall now supports automated allow-list generation

Amazon Route 53 Resolver DNS Firewall now automatically generates domain allow-lists from your VPCs’ observed DNS query traffic, helping you enforce least-privilege DNS access and reduce the risk of DNS-based data exfiltration.\nPreviously, implementing a least-privilege model required security teams to manually analyze DNS query logs to identify the domains their VPCs depend on before deploying allow-list rules. This created operational overhead and delayed enforcement of a deny-by-default (walled-garden) posture. With automated allow-list generation, DNS Firewall analyzes DNS query traffic using a lookback window configurable from 1 hour to 7 days to create an allow-list of observed domains. It then continuously monitors DNS queries and flags newly observed domains, so you can review them before adding them to the allow-list. You can deploy the rule across multiple VPCs and accounts using AWS Firewall Manager or Route 53 Profiles. This capability is available at no additional cost to DNS Firewall Advanced customers, in all AWS Regions where Route 53 DNS Firewall is supported, except in AWS GovCloud (US) and Amazon Web Services China Regions. To get started, use the Route 53 DNS Firewall console, AWS CLI, or API. To learn more, visit the Amazon Route 53 Resolver DNS Firewall product page and the Route 53 DNS Firewall Developer Guide.

Amazon EKS Auto Mode now rightsizes Pods to reduce compute costs (Preview)

Amazon Elastic Kubernetes Service (Amazon EKS) introduces resource optimization for EKS Auto Mode, a new capability that rightsizes the CPU and memory requests of running Pods based on their actual usage. CPU and memory requests are often set at deploy time and rarely revisited, and Auto Mode provisions EC2 instances to fit them. Rightsizing lets Auto Mode run the same Deployments and StatefulSets on fewer or smaller instances, without hand tuning and at no additional charge.\nTurn on resource optimization by creating a ClusterResourceOptimizationPolicy in your EKS cluster, review its recommendations in the default Recommend mode, and switch to Apply mode when you’re ready to resize running Pods in place, without recreating them. For example, a platform team running Java services under Argo CD can rightsize every application namespace without editing templates or creating GitOps drift, while resource optimization keeps memory above each Java heap and raises memory after an out-of-memory kill. It doesn’t change the CPU or memory that a Horizontal Pod Autoscaler or KEDA scales on, and you can exclude workloads that the Vertical Pod Autoscaler already manages. You can use resource optimization on EKS Auto Mode clusters running Kubernetes 1.36 or later in the following AWS Regions: US East (N. Virginia), US East (Ohio), US West (Oregon), Europe (Frankfurt), Europe (Ireland), Asia Pacific (Seoul), and Asia Pacific (Tokyo). To learn more, see Rightsize workloads with resource optimization in the Amazon EKS User Guide and the EKS Auto Mode product page.

AWS Split Cost Allocation Data now supports Amazon S3

You can now allocate the storage and request costs of a shared Amazon S3 general purpose bucket to its prefixes in the AWS Cost and Usage Report (CUR). Many organizations store data for multiple teams, applications, or tenants in one bucket under different prefixes such as “marketing/” and “finance/”. Previously, splitting that bucket’s bill by prefix required custom pipelines and your own pricing logic.\nWith this new capability, S3 Storage Lens publishes prefix-level storage and request metrics to AWS Billing and Cost Management. Billing and Cost Management then allocates the bucket’s storage and Tier 1 and Tier 2 request costs to the deepest matching prefix in the range of prefix levels you choose. Each prefix line item carries its full prefix path as AWS-generated cost allocation tags, so you can total costs by top-level prefix, team, or application, and reconcile it back to the bucket. You can use this data for chargeback and showback, and to see which prefixes drive your S3 spend. To get started, enable AWS Billing and Cost Management publishing in your S3 Storage Lens advanced tier configuration, then enable split cost allocation data for S3 in the AWS Billing and Cost Management console. This feature is available in all AWS Regions, except the AWS China Regions and the AWS GovCloud (US) Regions, at no additional charge beyond S3 Storage Lens advanced tier pricing. To learn more, visit the AWS Cost and Usage Reports product page, Understanding split cost allocation data, and Estimating Amazon S3 costs by prefix with S3 Storage Lens.

AWS Transit Gateway now shares resource context with firewalls for segmentation policies

AWS Transit Gateway now supports context sharing with security services, such as attached firewalls, allowing you to use Amazon Virtual Private Clouds (VPCs), VPC subnets, VPC security groups, and transit gateway attachments in firewall policies alongside IP addresses. Until now, attached firewalls segmented traffic by IP address, which is hard to plan and maintain, or by resource tags, which firewalls periodically synchronize with IP addresses and can miss new workloads in between. With this feature, you can enforce segmentation based on how workloads and attachments are defined in AWS.\nTransit Gateway embeds resource context in the traffic it sends to your firewall, which can enforce the right policy from the first packet. For example, you can apply threat protection to traffic between web-tier and database-tier security groups, block development VPCs from reaching production VPCs, apply a common inspection policy to traffic entering AWS through AWS Direct Connect and AWS Site-to-Site VPN attachments, or apply policies to short-lived AWS Lambda functions as soon as they send traffic, all without updating rules as IP addresses change. AWS Network Firewall and Palo Alto Networks Software Firewalls support policies using resource context alongside IP-based rules. To get started, create a context store in AWS Network Manager and associate your transit gateways. The context store monitors resources connected to those transit gateways and shares their context with firewalls. Context sharing with security services is available in all commercial AWS Regions where AWS Transit Gateway is available, except Middle East (Bahrain) and Middle East (UAE). It is also available in the AWS GovCloud (US-East) and AWS GovCloud (US-West) Regions. To learn more, visit the AWS Transit Gateway product page and context sharing with security services documentation.

AWS Lambda durable functions now supports pausing and deleting executions

AWS Lambda durable functions now lets you pause a running durable execution and resume it later without losing its progress. When you resume an execution, it continues from its last checkpoint on the same function version. By default, time spent paused doesn’t count toward the execution timeout. You can also now delete a completed execution’s data before its retention period ends.\nWith Lambda durable functions, you can build resilient multi-step applications and AI workflows that run for up to one year, with built-in checkpointing and automatic retries. With this launch, when a downstream service is degraded or a model provider’s rate limit is reached, you can pause affected executions while you resolve the issue instead of letting the retries run out. You can also delete an execution’s history, checkpoint data, and input and output payloads after it succeeds, fails, stops, or times out. For example, you can delete an execution whose input contains personal data, or clean up failed executions that a bug left behind. You can pause, resume, and delete executions using the Lambda console, the AWS CLI, or the AWS SDKs in all AWS Regions where AWS Lambda durable functions is available. To learn more, see Managing durable executions in the AWS Lambda Developer Guide, or visit the AWS Lambda durable functions product page. If you build with an AI coding assistant such as Kiro, the Agent Toolkit for AWS gives it skills and guidance for building on AWS.

AWS Workload Credentials Provider adds container support for ACM certificates

The open source AWS Workload Credentials Provider can now export AWS Certificate Manager (ACM) certificates to applications running in containers on Amazon Elastic Container Service (Amazon ECS), including AWS Fargate, Amazon Elastic Kubernetes Service (Amazon EKS), and Docker on Amazon EC2. You can now automate certificate deployment and renewal for containerized TLS workloads the same way you do on EC2 instances, on both Linux and Windows.\nThe provider runs as a sidecar in your ECS task or EKS pod. It exports each certificate and private key to a shared volume and writes the renewed files when ACM rotates the certificate. For TLS servers that don’t reload certificate files on their own, such as NGINX, Apache, HAProxy, or IIS, a companion watcher in the TLS container runs your reload command after each renewal. The watcher needs no AWS credentials. TLS servers that reload automatically, such as Envoy, Traefik, and Caddy, need only the shared volume. Access is controlled through IAM: the sidecar assumes a per-workload role that is allowed to export only that workload’s certificates. To get started, download the provider and watcher binaries for Linux (x86_64 and Arm64) or Windows, and follow the setup steps and ECS and EKS examples in the AWS Workload Credentials Provider README on GitHub. The provider is available at no additional charge, and standard ACM charges apply to exported certificates. To learn more, see the AWS Certificate Manager documentation.

Amazon S3 Storage Lens now tracks S3 Intelligent-Tiering access tiers

Amazon S3 Storage Lens now provides storage metrics for each S3 Intelligent-Tiering access tier. You can see object counts and storage bytes in every tier across your buckets. This helps you track how your data moves between tiers, estimate storage costs, and measure your savings.\n S3 Storage Lens gives you visibility across your organization into storage usage and activity to help you optimize costs and performance. Access tier metrics appear in your S3 Storage Lens dashboards and metrics exports alongside your existing metrics. S3 Storage Lens also now publishes prefix metrics to AWS Billing and Cost Management, so you can allocate shared bucket costs to teams and applications with split cost allocation data for S3.

To get started, enable access tier metrics in your S3 Storage Lens advanced tier configuration. Access tier metrics are available in all AWS Regions, except the AWS China Regions and the AWS GovCloud (US) Regions. To learn more, visit the S3 Storage Lens documentation. For pricing, visit the S3 pricing page.

AWS Blogs

AWS Japan Blog (Japanese)

AWS News Blog

AWS Architecture Blog

AWS Cloud Financial Management

AWS Cloud Operations Blog

AWS Big Data Blog

AWS Compute Blog

Containers

AWS Database Blog

Desktop and Application Streaming

AWS Developer Tools Blog

AWS DevOps & Developer Productivity Blog

AWS for Industries

Artificial Intelligence

AWS for M&E Blog

AWS Quantum Technologies Blog

AWS Security Blog

AWS Storage Blog

Open Source Project

AWS CLI

Amplify for iOS

Amazon Chime SDK for JavaScript