You can now launch Apache ActiveMQ 5.15.14 brokers on Amazon MQ. This patch update to ActiveMQ contains several fixes and new features compared to the previously supported version, ActiveMQ 5.15.13.

Today, we are excited to release a new learning capsule that dives deep into Transformer models. To learn the concepts behind generative AI, developers can use easy-to-consume, bite-sized learning capsules in the AWS DeepComposer console. In this new learning capsule , developers will learn about Transformer models which use attention layers to capture dependencies non-sequentially. Transformers are a machine learning architecture model designed to handle sequential data. By building a transformer-based model that understands musical motifs, or sequences, developers can create a model that can generate new music based on a series of sequences learned during training. Get started with creating a composition using the Transformers technique in the AWS DeepComposer Music studio .

AWS Systems Manager Automation now lets you define the actions to perform when an operator cancels the execution of a runbook. Using this feature, you can automatically roll back any changes or perform necessary cleanups when the execution of a runbook is canceled.

You can now reboot individual brokers within your Amazon Managed Streaming for Apache Kafka (Amazon MSK) cluster by using the Reboot Broker API. The Reboot Broker API enables you to test the resiliency of your data producers and consumers by simulating MSK service maintenance such as automated patching or the effect a version upgrade has on a MSK broker. Visit the Amazon MSK API Reference for information on how to use this API. 

Amazon Managed Blockchain now supports Ethereum in preview. Ethereum is a popular decentralized blockchain platform that establishes a peer-to-peer network allowing participants to transact without a trusted central authority. AWS customers can easily provision Ethereum nodes in minutes, and connect to the public Ethereum main network and test networks such as Rinkeby and Ropsten. With Amazon Managed Blockchain, customers get secure networking, fast and reliable syncs to the Ethereum blockchain, durable elastic storage for ledger data, encryption at rest and transport, and secure access to the network via standard open-source Ethereum APIs. Amazon Managed Blockchain automatically scales the storage, monitors node health, replaces unhealthy nodes, and automates Ethereum software upgrades, improving the availability of customers’ Ethereum infrastructure.

Amazon AppFlow, a fully managed data integration service that enables customers to securely transfer data between AWS services and cloud applications, added Upsolver as a destination. You can now use Amazon AppFlow to configure data flows with Upsolver as the destination. Upsolver is a data lake ETL platform that stores data on the cloud, applies advanced transformations, and distributes it to multiple destinations.

The AWS Load Balancer Controller provides a Kubernetes native way to configure and manage Elastic Load Balancers that route traffic to applications running in Kubernetes clusters. Elastic Load Balancing offers multiple load balancers that all feature the high availability, automatic scaling, and robust security necessary to help make your applications fault tolerant.

You can now connect your on-premises AWS Outposts infrastructure privately over Direct Connect to services running in the AWS region using AWS Outposts Private Connectivity.  

The EC2 Hibernation feature is now available in the AWS GovCloud (US) Regions. Hibernation gives you the ability to launch EC2 instances, set them up as desired, hibernate them, and then quickly bring them back to life when you need them. Applications pick up exactly where they left off instead of rebuilding their memory footprint. By using Hibernation, you can maintain a fleet of pre-warmed instances to get to a productive state faster without modifying your existing applications.

Amazon Transcribe Medical is a HIPAA-eligible automatic speech recognition (ASR) service that makes it easy for developers to add speech-to-text capabilities to their healthcare and life science applications. Starting today, you can use the new Channel Identification feature to transcribe the speech from both live audio streams or static audio files that contain multiple audio channels. With Channel Identification, you can process audio from multiple channel to generate a single transcript of the conversation complete with channel labels.