11/18/2026, 12:00:00 AM ~ 11/19/2026, 12:00:00 AM (UTC)

Recent Announcements

Amazon OpenSearch Memory is now generally available

Today, AWS announced the general availability of Amazon OpenSearch Memory, a new serverless capability in Amazon OpenSearch Service for adding persistent memory to AI agents. OpenSearch Memory stores agent interactions as searchable events, derives useful memories, and retrieves relevant context across sessions, so agents can recall prior decisions, earlier interactions, and user preferences without replaying the full conversation history to the model. Customers can give agents long-term memory without provisioning infrastructure or building separate extraction, indexing, consolidation, and retrieval systems.\n OpenSearch Memory stores the original events verbatim, makes them searchable within seconds, and asynchronously derives structured memories. It manages the full memory lifecycle, including extraction, embedding, indexing, consolidation (adding, updating, or skipping memories as new information arrives), and retrieval using keyword, semantic, and hybrid search. Customers control access with AWS Identity and Access Management (IAM) for each memory store, with actor and session scoping, and can set expiration periods for events and memories. Because it is serverless, there are no clusters to provision, instance types to choose, or indexes to create, and capacity for extraction and search scales automatically.

Amazon OpenSearch Memory is available today in US East (N. Virginia, Ohio), US West (Oregon), Europe (Frankfurt), and Asia Pacific (Tokyo) Regions. For pricing details about Amazon OpenSearch Memory, visit the pricing page.

To learn more about Amazon OpenSearch Memory, see the technical documentation and launch blog. You can get started by visiting the technical launch blog that details how to add persistent memory to your agents with Amazon OpenSearch Memory.

Introducing the Apache Spark Optimization Agent for AWS Glue ETL

AWS Glue ETL now offers the Apache Spark Optimization Agent, which analyzes completed Spark job runs and delivers cost and performance optimization recommendations within minutes. You can select a cost or performance goal, and the agent returns specific configuration changes with the telemetry evidence and reasoning behind each suggestion. You can use the agent on any completed job run from the AWS Glue console, or from your own tools through the Amazon SageMaker Unified Studio MCP Server.\n The agent takes the job’s configuration, logs, and Spark telemetry together with your chosen goal 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. When cost and performance trade off, the agent improves the metric you chose while keeping any regression of the other within defined limits. The engine also checks best practices such as platform-level improvements like upgrading to a newer AWS Glue version or using AWS Graviton processors. Each recommendation is labeled by the area it modifies: infrastructure, Spark configuration, code, or data layout. All access is authenticated with AWS Identity and Access Management (IAM) and logged in AWS CloudTrail.

The Apache Spark Optimization Agent is available at no additional charge for AWS Glue versions 4.0 and later in all Regions where AWS Glue is available, except the AWS GovCloud (US) and China Regions. To get started, see Apache Spark Optimization Agent for Amazon EMR and AWS Glue and the blog post - Introducing the Apache Spark optimization agent for Amazon EMR Serverless and AWS Glue ETL.

Amazon SageMaker AI now supports cross-Region inference routing

Today, Amazon SageMaker AI announces cross-Region inference routing, which helps keep inference applications available when a region runs short of capacity by sending requests to endpoints deployed in other AWS Regions and AWS accounts. Models served on accelerated instances can reach capacity limits in a single Region, and an unhealthy endpoint can lead to throttling and errors for end users. Previously, teams had to build and maintain their own routing, health checks, and retry logic across Regions.\n With cross-Region inference routing in SageMaker AI, developers create an inference router, register the real-time endpoints, asynchronous endpoints, and inference components they already run, and point their applications at the router. The router prefers endpoints in its own Region, shifts new requests to healthy destinations in other Regions as errors rise, and returns traffic gradually after the local endpoint recovers. Existing endpoints, models, and containers do not change. Endpoint owners stay in control: an endpoint policy determines which routers in other accounts can use an endpoint, and an optional cross-Region traffic cap reserves capacity for the endpoint’s local callers. Each real-time response identifies the endpoint that served the request and the reason, and router metrics in Amazon CloudWatch show where traffic went. Cross-Region inference routing is available for real-time inference in 32 commercial AWS Regions, and for asynchronous inference in 25 Regions. There is no additional charge for cross-Region inference routing. Customers pay standard SageMaker AI pricing for the endpoints that serve requests and standard Amazon CloudWatch rates for router metrics. To learn more, see the Amazon SageMaker AI Developer Guide and SageMaker AI Deployment page.

Amazon CloudFront standard logging now supports CMCD and additional log fields

Amazon CloudFront standard logging now supports additional access log fields that were previously available only in CloudFront real-time logs. You can now include these fields in your standard logs and deliver them to Amazon S3, Amazon CloudWatch Logs, or Amazon Data Firehose, without setting up an additional data stream for real-time logs.\n The new fields include the IP address of the CloudFront server that handled the request, identifiers that separate traffic on a primary distribution from traffic on a staging distribution, and a media quality confidence score. They also include Common Media Client Data (CMCD) fields, which video players send with each request to report details like bitrate, buffer length, and playback session ID. Streaming teams can use these fields to match playback problems that viewers see with CloudFront delivery data, which makes troubleshooting faster.

These fields are available in all CloudFront edge locations, and standard logging pricing applies. To learn more, see Standard logging page in the Amazon CloudFront Developer Guide.

Amazon CloudFront Announces WebSocket and gRPC Support for Origin Mutual TLS

Amazon CloudFront (CF) now supports WebSocket and gRPC traffic to application origins in AWS regions, with mutual TLS authentication between CF and the application origin (origin mTLS). Customers can now use TLS certificates to authenticate requests from CloudFront while setting up WebSocket or gRPC connections to their origins. Origin mTLS protects internet facing applications that need persistent, bidirectional or streaming connections, such as chat platforms, collaborative editing tools or live dashboards from unauthenticated requests.\n Previously, customers serving WebSocket or gRPC traffic through CloudFront had to use shared secrets or IP allow-lists to identify requests originating from CloudFront. This was operationally complex and did not provide strong authentication. With Origin mTLS, customers can use one TLS based authentication for HTTP, WebSocket and gRPC traffic to an application origin. This works with application origins within AWS such as Application Load Balancer and API Gateway, on-premises origins, and origins hosted outside AWS.

There is no additional charge for origin mTLS, and it is available with pay-as-you-go pricing and in the Business and Premium flat-rate pricing plans. To learn more, visit the CloudFront origin mutual TLS documentation.

AWS Lambda adds support for GPU instances on Lambda Managed Instances

AWS Lambda now supports GPU-accelerated Event Functions on Lambda Managed Instances using EC2 G-series instances. You can now run GPU-accelerated workloads including AI inference, media processing, 3D rendering, scientific computing, and financial simulations on Lambda without managing GPU drivers, frameworks, or scaling. Lambda runs multiple invocations concurrently in each execution environment and scales responsively based on GPU utilization, CPU utilization, and concurrency to keep your GPUs busy and reduce costs. GPU-accelerated functions integrate with the same 200+ AWS event sources as other Event Functions, so you can add GPU compute as another step in your serverless application.\n Running GPU workloads today requires customers to provision hardware, manage driver and framework dependencies, and build custom scaling and capacity management. With this launch, you can go from code to a running GPU workload in minutes. Lambda provides GPU base images preconfigured with frameworks such as vLLM, SGLang, and PyTorch. You can run workloads with large artifacts using container images and NVMe ephemeral storage of up to 32 GB each. To reduce startup time, Lambda supports Seekable OCI (SOCI) lazy loading and streaming artifacts directly from Amazon S3. For resilience, you can select G-series instance types from multiple generations in the allowed instance types for your capacity provider. Lambda provisions instances from the newest generation that meets your function requirements and falls back to older generations when capacity is unavailable.

GPU-accelerated Event Functions on Lambda Managed Instances is available in all AWS Commercial Regions where Lambda Managed Instances and EC2 G-series instances are available. You can configure GPU-accelerated functions using the AWS Console, AWS CLI, AWS SDKs, AWS CloudFormation, AWS CDK, or AWS SAM. You can also use this feature through coding assistants such as Kiro, Claude Code, or Codex by installing the Lambda Managed Instances skill from the Agent Toolkit for AWS. You pay the EC2 On-Demand price (Savings Plans and Reserved Instances apply) for your instance, plus a 9.75% Lambda management fee and $0.20 per million requests. To learn more, visit the AWS Lambda Developer Guide and AWS Lambda Pricing.

Amazon RDS for MySQL supports MySQL minor version 8.4.13

Starting today, Amazon Relational Database Service (Amazon RDS) for MySQL supports MySQL minor version 8.4.13, the latest minor released by community MySQL. We recommend upgrading to the newer minor versions to accept fixes for Common Vulnerabilities and Exposures (CVEs) in prior versions of MySQL and to benefit from bug fixes, performance improvements, and new functionality added by the MySQL community. Learn more about the enhancements in RDS for MySQL 8.4.13 in the Amazon RDS user guide and MySQL 8.4.13 release notes.\n You can upgrade your databases during scheduled maintenance windows using automatic minor version upgrades. To simplify operations at scale, enable automatic minor version upgrades and use the AWS Organizations Upgrade Rollout Policy to orchestrate upgrades across your instances in phases, validating on lower-priority environments before upgrading your most critical ones. You can perform minor version upgrades in-place or via snapshot restore. Learn more about performing version upgrades in the Amazon RDS User Guide. This release is supported in all AWS Regions where Amazon RDS for MySQL is available.

Amazon RDS for MySQL makes it simple to set up, operate, and scale MySQL deployments in the cloud. Create or update a fully managed Amazon RDS database in the Amazon RDS Management Console or by using the AWS Command Line Interface (AWS CLI).

Amazon Cognito now supports point-in-time recovery

Amazon Cognito now supports point-in-time recovery (PITR), so you can restore the user data in a user pool to an earlier time. Amazon Cognito helps you implement secure sign-in and access control for users, AI agents, and microservices. Previously, you had to build and maintain your own export and restore process, and restored users received new identifiers and had to reset their passwords.\n With PITR, you choose a recovery point, preview how many users, groups, multi-factor authentication (MFA) settings, and passkeys the recovery will create, update, or delete, and then start or cancel the recovery. Recovered users keep their identifiers, passwords, MFA settings, and passkeys, so they can sign in without resetting their credentials. For example, if a script deletes or modifies users by mistake, you can restore the user data to the moment before it ran.

PITR is enabled automatically for user pools on the Essentials and Plus feature plans at no additional charge. The recovery window is up to 1 day on Essentials and up to 14 days on Plus. You can initiate and manage a recovery using the AWS Management Console, the Amazon Cognito API, AWS Command Line Interface (AWS CLI), or AWS SDKs. PITR is available in all AWS Regions where Amazon Cognito is available. To learn more, see Point-in-time recovery in the Amazon Cognito Developer Guide.

Amazon RDS for MySQL now supports MySQL major version 9.7

Starting today, Amazon RDS for MySQL supports MySQL major version 9.7, the latest Long-Term Support release from MySQL community. Amazon RDS for MySQL 9.7 launches with community MySQL 9.7.4. This launch introduces support for the Replication Applier Metrics component, JSON Duality Views, and Hypergraph Optimizer.  \n The Replication Applier Metrics help you diagnose replication lag by adding two Performance Schema tables that report per-channel applier backlog, throughput, and wait statistics, along with each worker’s progress on its current transaction. JSON Duality Views let applications read and write the same relational tables as JSON documents. The Hypergraph Optimizer is an opt-in query optimizer that can improve execution plans for complex, multi-join queries. For more details, refer to the MySQL 9.7.4 release notes and the Amazon RDS for MySQL release notes.

You can upgrade your database using Amazon RDS Blue/Green Deployments, in-place upgrade, or restore from a snapshot. Learn more about performing version upgrades in the Amazon RDS User Guide. This release is supported in all AWS Regions where Amazon RDS for MySQL is available. Learn more about pricing details at Amazon RDS for MySQL pricing.

Amazon RDS for MySQL makes it simple to set up, operate, and scale MySQL deployments in the cloud. Create or update a fully managed Amazon RDS database in the Amazon RDS Management Console or by using the AWS Command Line Interface (AWS CLI).

Amazon EKS Hybrid Nodes reduces price for large hosts

Amazon Elastic Kubernetes Service (Amazon EKS) Hybrid Nodes now has lower pricing for large nodes, capping the billed vCPU per node at 32 vCPU, so you are no longer charged for vCPUs above that per node threshold. Amazon EKS Hybrid Nodes lets you use your existing on-premises and edge infrastructure as nodes in Amazon EKS clusters. You can now use your largest servers, including those with GPUs, with Amazon EKS Hybrid Nodes at lower cost.\n With this change, the Amazon EKS Hybrid Nodes cost of a large server no longer grows with its vCPU count. At the first pricing tier, a 64 vCPU node now costs 50% less than before, and a 256 vCPU node costs 88% less, as both are now billed only for 32 vCPUs. This makes Amazon EKS Hybrid Nodes more cost-effective for large bare-metal and GPU servers running compute-intensive workloads, such as AI/ML training and inference, that must stay on-premises for data residency, latency, or regulatory reasons. The pricing change applies automatically to all new hybrid nodes. For the new pricing to take effect for existing hybrid nodes attached to clusters, delete and reattach them. Nodes with 32 vCPUs or fewer are still billed for their actual vCPUs, and existing tiered pricing still applies, so your effective rate falls as monthly usage grows.

The new pricing is available in all AWS Regions where Amazon EKS Hybrid Nodes is available. To learn more about getting started and updated pricing details, visit the Amazon EKS Hybrid Nodes documentation and Amazon EKS pricing page.

Amazon CloudFront Functions now supports network calls

Amazon CloudFront Functions can now make network calls and act on the response during a request. CloudFront Functions runs lightweight JavaScript at CloudFront edge locations to customize requests and responses. Previously, if your logic depended on data from another service, you had to use Lambda@Edge or move that logic to your origin. Now your functions can use real-time data to personalize, authenticate, or route requests at the edge, close to your viewers.\n With the new cf.fetch() method, your functions can call your API endpoints and use the response to decide how to handle each request. For example, you can check an AI agent’s permissions and usage budget so that only authorized requests reach your API or MCP server, or send request signals to a security or fraud detection service to block or challenge bad traffic. With cf.beacon(), you can send analytics events in the background without slowing down viewers, even for content served from cache. Both methods take advantage of the origin settings already on your distribution, including origin access control (OAC), VPC origins, and connection reuse.

Network calls are available at all CloudFront edge locations. They cost $0.10 per million calls, on top of CloudFront Functions invocation charges, and are included at no additional cost in the Pro plan and above in CloudFront flat-rate plans. To learn more, see Helper methods for network calls in the Amazon CloudFront Developer Guide and Amazon CloudFront pricing.

Amazon RDS now supports fast storage volume initialization

Amazon RDS now initializes storage volumes faster when you perform a point-in-time restore (PITR), create a read replica, or convert a Single-AZ instance to Multi-AZ. Each of these operations creates storage volumes from a snapshot, and the data blocks are downloaded from Amazon S3 before you can access them. With fast storage volume initialization, you get a fully performant DB instance in a predictable amount of time.\n Previously, the initialization rate varied with your workload and with which blocks were accessed. As there was no dedicated bandwidth, so you could see increased I/O latency and had no way to tell when an instance would reach full performance. Now, each database instance is initialized at up to 4 TiB per hour, across every gp3 and io2 storage volume attached to the instance. If concurrent initialization across your account exceeds the available bandwidth, the remaining blocks initialize on a best-effort basis, so initialization is never slower than today. Your DB instance stays available and usable while initialization runs.

Fast storage volume initialization is enabled by default and no action is required. To follow progress, use the StorageOperationStatus and StorageOperationPercentProgress fields in the Amazon RDS console, the DescribeDBInstances API, the AWS Command Line Interface (CLI), or AWS SDKs.

This capability is available at no additional cost for RDS for MariaDB, RDS for MySQL, RDS for PostgreSQL, RDS for Db2, RDS for Microsoft SQL Server, and RDS for Oracle, in all commercial AWS Regions and the AWS GovCloud (US) Regions. To learn more, see Amazon RDS storage in the Amazon RDS User Guide.

AWS Clean Rooms supports writing PySpark job results to intermediate tables

AWS Clean Rooms now supports intermediate tables for PySpark jobs, offering increased flexibility for organizations running complex, multi-step analytical workflows with their partners. With this launch, customers can write the results of a PySpark job to an intermediate table for reuse in subsequent analyses. For example, a measurement vendor can perform feature engineering with PySpark, then use those results as input into an ML model training job to identify new high-value customers to inform audience targeting and bid optimization for their campaigns.  \n AWS Clean Rooms helps companies and their partners easily analyze and collaborate on their collective datasets without revealing or copying one another’s underlying data. For more information about the AWS Regions where AWS Clean Rooms is available, see the AWS Regions table. To learn more about collaborating with AWS Clean Rooms, visit AWS Clean Rooms.

AWS Lambda now supports customer-managed runtimes, giving platform teams centralized control over runtime contents and updates

AWS Lambda now supports customer-managed runtimes for Lambda Event Functions (formerly Lambda functions). Customer-managed runtimes are a standalone resource type that fully decouples runtime management from function deployment. Platform teams gain centralized control over the runtime contents, updates, access, and lifecycle, including the ability to push updates to shared components without per-function redeployments. Application teams consume the runtime from their functions in the same way as a Lambda-managed runtime, preserving their familiar developer experience.\n Many organizations require every function to include approved components, such as SDKs, certificates, and security or monitoring tools. Previously, sharing these through Lambda Layers or container images meant redeploying every function to pick up a change. With customer-managed runtimes, platform teams manage these components in a central runtime, packaged as OCI container images in Amazon ECR and compatible with existing Docker and CI/CD tooling. Functions pick up updates automatically as Lambda recycles execution environments, so there’s no need to redeploy each function or coordinate rollouts. Updates can be scheduled and staged to match change control policies, with centralized health metrics and rapid rollback. With this launch, the existing custom runtimes feature is renamed “bring your own language.”

AWS Lambda customer-managed runtimes is now generally available in all commercial AWS Regions. There is no additional charge for customer-managed runtimes; functions are billed at standard Lambda rates and standard Amazon ECR pricing applies to runtime images.

To get started, build your runtime as a container image in Amazon ECR, starting from a Lambda base image or from scratch, then create a runtime, a version, and an alias. Application teams reference the runtime alias in their function configuration and keep deploying their code as .zip files. Share runtimes with other accounts or an AWS Organization by using a resource-based policy. You can create and manage customer-managed runtimes using the AWS CLI, AWS SDKs, CloudFormation, or AWS CDK. To learn more, see the customer-managed runtimes documentation and launch blog post.

Amazon Aurora DSQL now supports audit logging

Amazon Aurora DSQL now supports database-internal audit logging, a fully managed capability that captures SQL statements executed on your cluster. Previously, AWS CloudTrail recorded who connected to a DSQL cluster, but not what they did once inside it. Now, financial institutions and regulated enterprises can meet production compliance requirements by capturing DDL, DML, DCL, and SELECT statements with full statement text, bind parameters, acting principal (both PostgreSQL role and AWS IAM identity), timestamp, table name, and transaction identifier.\n With audit logging, you can feed SIEM systems directly from Amazon S3, satisfy auditor requirements for change accountability, and reconstruct events after security incidents. Logs are delivered as structured JSON to your S3 bucket via CloudWatch Vended Logs. Configuration requires a single SQL setting—no extension to load and no log-shipping pipeline to build.

Audit logging is available in all AWS Regions where Aurora DSQL is available. Audit logging will incur Aurora DSQL DPU charges, charges for log delivery through CloudWatch Logs, and charges for storing audit records in Amazon S3. For pricing information, see Aurora DSQL pricing. To learn more, see audit logging in the Aurora DSQL User Guide.

AWS Control Tower now supports existing Amazon S3 bucket for log delivery

AWS Control Tower now lets you use an existing Amazon S3 bucket in your landing zone to store AWS Config and AWS CloudTrail logs. Previously, AWS Control Tower required you to use the Amazon S3 bucket it creates and manages for log storage. Now, customers who want full control of bucket policies, lifecycle rules, and logging infrastructure can consolidate centralized logs in an Amazon S3 bucket they already manage.\nYou can use your own Amazon S3 bucket for both AWS Config and AWS CloudTrail, giving you full control of bucket policies, lifecycle rules, and logging infrastructure. You can switch between self-managed and AWS Control Tower managed buckets at any time, and your Amazon S3 buckets are preserved during landing zone updates, resets, and decommissioning. The bucket must be owned by an AWS account in your organization, be located in your AWS Control Tower home Region, and be preconfigured with the bucket policy that AWS Control Tower requires. This feature is supported in landing zone version 4.0 and later and is available in all commercial AWS Regions, the AWS GovCloud (US) Regions, and the AWS European Sovereign Cloud where AWS Control Tower is available. To learn more about configuring self-managed Amazon S3 bucket for your AWS Control Tower landing zone, visit the AWS Control Tower documentation.

AWS IAM account access manager is now even easier to deploy and use

Deploying and using AWS Identity and Access Management (IAM) account access manager is now even easier: you can use AWS managed policies to enable and administer it; you can include it in your infrastructure-as-code (IaC) configurations; and its role assignments are integrated in a growing number of third-party identity and entitlement management solutions.\n Administering workforce access with account access manager requires permissions across several services, including AWS IAM Identity Center, AWS Organizations, AWS KMS, and AWS IAM. Previously, defining administrative access required hand-assembling multi-service IAM policies. The new AWS managed policies enable faster ramp up with broadly scoped job-aligned permissions for: 1/ Full administrator (AWSAccountAccessManagerAdministrator) - used to enable and manage account access manager in the AWS organization management account and to map users/groups to IAM roles; 2/ Entitlement Manager (AWSAccountAccessManagerEntitlementManager) - used to assign and remove user access without the ability to change service configuration; and 3/ ReadOnly (AWSAccountAccessManagerReadOnlyAccess) - used by auditors and compliance reviewers for visibility into user access.

Account access manager streamlines assignment of IAM roles to workforce users and groups in IAM Identity Center, bringing together access management flexibility, user awareness, and a single point of federation. You can include account access manager in your IaC configurations managed through Terraform, Pulumi, or AWS CloudFormation. The entitlements that you assign with account access manager are reflected in your identity provider (IdP) and entitlements management system, including those offered by AWS partners Okta, PingIdentity, Cisco Duo, Apono, BeyondTrust, Britive, Sonrai Security, C1.ai, Lumos, and SailPoint.

IAM account access manager is provided at no additional cost and available in all AWS Commercial Regions enabled by default. To learn more and get started, visit the AWS Identity and Access Management User Guide.

Amazon S3 Replication adds exclusions and Multi-Region key resolution

You can now replicate data in Amazon S3 buckets with fewer replication rules using two new S3 Replication capabilities, exclusions and Multi-Region key resolution. Buckets that many tenants or applications share often need many rules to skip certain data or to match encryption keys in each AWS Region. With these capabilities, customers who manage multi-tenant and shared buckets can replace many replication rules with a single rule.\n With exclusions, you can replicate an entire bucket while skipping specific prefixes or tagged objects, such as logs or experimental data. With Multi-Region key resolution, S3 Replication encrypts each replica with the matching AWS Key Management Service (AWS KMS) Multi-Region key in the destination Region, so replicas meet the same encryption requirements as their source objects. For example, a bucket that stores data for hundreds of tenants, each with its own Multi-Region key, can replicate with a single rule, including tenants you add later.

Exclusions and Multi-Region key resolution are optional and available at no additional charge in all AWS Regions. To get started, use the AWS SDKs, AWS CLI, S3 API, or S3 console. To learn more, visit the S3 Replication product page and Replicating objects within and across Regions in the S3 User Guide.

Amazon SageMaker HyperPod now supports higher pod density on EKS-orchestrated clusters

Amazon SageMaker HyperPod now supports higher pod density per node on EKS-orchestrated clusters, allowing you to get more out of every instance. This leaves more room for your applications alongside system DaemonSets for monitoring, logging, and security, and helps increase GPU, CPU and memory utilization on nodes, allowing you to run the same work using fewer instances.\n As you schedule more pods, the Amazon VPC CNI automatically adds network capacity to the node. Autoscaling with Karpenter also supports the higher pod capacity, so new nodes take on more pods as your cluster scales out.

Here are a few examples of workloads that benefit from higher pod density:

Model inference - Serve more model replicas or more models from each GPU instance, including on smaller types such as ml.g5 and ml.g6.

Interactive development - Host more notebooks, IDEs, and data science environments on each node.

GPU partitioning - Use all the Multi-Instance GPU (MIG) slices on instances such as ml.p5en.48xlarge, where each slice runs its own pod.

Operators and Controllers - Run operators, controllers, and observability agents on CPU nodes alongside your workloads, without crowding them out.

This feature is available for HyperPod EKS-orchestrated clusters in all AWS Regions where SageMaker HyperPod is available. To get started, update your cluster’s IAM permissions, cluster software, and Amazon VPC CNI add-on. To learn more, please check the documentation.

AWS Blogs

AWS Japan Blog (Japanese)

AWS News Blog

AWS Architecture Blog

AWS Cloud Operations Blog

AWS Big Data Blog

AWS Compute Blog

AWS Contact Center

Containers

AWS Database Blog

AWS Developer Tools Blog

AWS DevOps & Developer Productivity Blog

AWS for Industries

Artificial Intelligence

AWS for M&E Blog

Networking & Content Delivery

AWS Security Blog

AWS Storage Blog

Open Source Project

AWS CLI

OpenSearch

Amplify for iOS

Bottlerocket OS

Amazon EKS Anywhere