Relevant News for Germany
Relevant News for GermanyHere are the latest news items for Germany.
Published: 2026-02-23 16:56:24+00:00
Link: https://aws.amazon.com/about-aws/whats-new/2026/03/aws-batch-configurable-scale-down-delay/
AWS Batch now allows you to configure a scale down delay for managed compute environments, helping reduce job processing delays for intermittent and periodic workloads. With the new minScaleDownDelayMinutes parameter, you can specify how long AWS Batch keeps instances running after their jobs complete (from 20 minutes to 1 week), preventing unnecessary instance terminations and relaunches that can delay subsequent job processing.
You can configure the scale down delay when creating or updating a compute environment via the AWS Batch API (CreateComputeEnvironment or UpdateComputeEnvironment) or the AWS Batch Management Console. The delay is applied at the instance level, based on when each instance last completed a job.
Scale down delay is supported today in all AWS Regions where AWS Batch is available. For more information, see the AWS Batch API Guide.
Published: 2026-03-02 19:05:00+00:00
Amazon Bedrock batch inference now supports the Converse API as a model invocation type, enabling you to use a consistent, model-agnostic input format for your batch workloads.
Previously, batch inference required model-specific request formats using the InvokeModel API. Now, when creating a batch inference job, you can select Converse as the model invocation type and structure your input data using the standard Converse API request format. Output for Converse batch jobs follows the Converse API response format. With this feature, you can use the same unified request format for both real-time and batch inference, simplifying prompt management and reducing the effort needed to switch between models. You can configure the Converse model invocation type through both the Amazon Bedrock console and the API.
This capability is available in all AWS Regions that support Amazon Bedrock batch inference. To get started, see Create a batch inference job and Format and upload your batch inference data in the Amazon Bedrock User Guide.
Published: 2026-02-27 19:00:00+00:00
Amazon OpenSearch Service has enhanced Cluster Insights with two new insights — Cluster Overload and Suboptimal Sharding Strategy. Suboptimal Sharding Strategy provides instant visibility into shard imbalances that cause uneven workload distribution, while Cluster Overload surfaces elevated cluster resource utilization that can lead to request throttling or rejections. Both insights come with details of affected resources along with actionable mitigation recommendations.
Previously, identifying resource constraints and shard imbalances required manually correlating multiple metrics and logs, making it difficult to detect issues early. With these new insights, you can proactively monitor cluster health and take timely action.
Suboptimal Sharding Strategy detects shard imbalances caused by indices with too few shards relative to the number of data nodes, or by shards carrying disproportionately large amounts of data compared to others. It identifies the root cause of uneven workload distribution and provides recommendations to help you achieve optimal shard distribution for improved query performance and resource utilization. Similarly, Cluster Overload helps you identify elevated resource utilization, including CPU, memory, disk I/O, disk throughput, and disk utilization that can potentially lead to request throttling or rejections. It also provides scale-up recommendations so you can take timely action to protect your critical workloads.
These new insights are available at no additional cost for OpenSearch version 2.17 or later in all Regions where the OpenSearch UI is available. See the complete list of supported Regions here. To learn more, visit the Cluster Insights documentation or view the complete catalog of available insights.
Published: 2026-02-27 10:49:00+00:00
Amazon Bedrock now supports OpenAI-compatible Projects API in the Mantle inference engine in Amazon Bedrock. Amazon Bedrock is a fully managed service that offers a broad selection of best-in-class foundation models from leading AI companies like Anthropic, Meta, and OpenAI, along with a broad set of specialized developer tools that make it easy to build and scale compelling generative AI applications. Mantle is Amazon Bedrock's distributed inference engine for large-scale model serving that supports OpenAI-compatible APIs.
With Projects API, customers who have more than one application, environment, or team can now create individual projects to achieve better isolation across all of them. You can assign different IAM-based access control to each project and add tags to each project for better cost visibility.
Projects are available for all customers using the OpenAI-compatible APIs, the Responses API and Chat Completions API, through the Mantle inference engine in Amazon Bedrock. There is no additional charge for using the Projects API. You pay only for the underlying model inference you consume. To get started with the Projects API in Amazon Bedrock, visit the Amazon Bedrock documentation.
Published: 2026-02-26 23:06:00+00:00
Link: https://aws.amazon.com/about-aws/whats-new/2026/02/lambda-durable-execution-java-preview/
Today, AWS announces the developer preview of the AWS Lambda Durable Execution SDK for Java. With this SDK, developers can build resilient multi-step applications like order processing pipelines, AI-assisted workflows, and human-in-the-loop approvals using Lambda durable functions, without implementing custom progress tracking or integrating external orchestration services.
Lambda durable functions extend Lambda's event-driven programming model with operations that checkpoint progress automatically and pause execution for up to a year when waiting on external events. The new Durable Execution SDK for Java provides an idiomatic experience for building with durable functions and is compatible with Java 17+. This preview includes steps for progress tracking, waits for efficient suspension, and durable futures for callback-based workflows.
To get started, see the Lambda durable functions developer guide and the AWS Lambda Durable Execution SDK for Java on GitHub. To learn more about Lambda durable functions, visit the product page.
On-demand functions are not billed for duration while paused. For pricing details, see AWS Lambda Pricing. For information about AWS Regions where Lambda durable functions are available, see the AWS Regional Services List.
Published: 2026-02-26 07:00:00+00:00
Link: https://aws.amazon.com/about-aws/whats-new/2026/02/aws-appconfig-new-relic-for-automated-rollback/
AWS AppConfig today launched a new integration that enables automated, intelligent rollbacks during feature flag and dynamic configuration deployments using New Relic Workflow Automation. Building on AWS AppConfig's third-party alert capability, this integration provides teams using New Relic with a solution to automatically detect degraded application health and trigger rollbacks in seconds, eliminating manual intervention.
When you deploy feature flags using AWS AppConfig's gradual deployment strategy, the AWS AppConfig New Relic Extension continuously monitors your application health against configured alert conditions. If issues are detected during a feature flag update and deployment, such as increased error rates or elevated latency, the New Relic Workflow automatically sends a notification to trigger an immediate rollback, reverting the feature flag to its previous state. This closed-loop automation reduces the time between detection and remediation from minutes to seconds, minimizing customer impact during failed deployments.
Published: 2026-02-24 16:00:00+00:00
Link: https://aws.amazon.com/about-aws/whats-new/2026/02/amazon-bedrock-reinforcement-fine-tuning-openai
Amazon Bedrock now extends reinforcement fine-tuning (RFT) support to popular open-weight models, including OpenAI GPT-OSS and Qwen models, and introduces OpenAI-compatible fine-tuning APIs. These capabilities make it easier for developers to improve open-weight model accuracy without requiring deep machine learning expertise or large volumes of labeled data. Reinforcement fine-tuning in Amazon Bedrock automates the end-to-end customization workflow, allowing models to learn from feedback on multiple possible responses using a small set of prompts, rather than traditional large training datasets. Reinforcement fine-tuning enables customers to use smaller, faster, and more cost-effective model variants while maintaining high quality.
Organizations often struggle to adapt foundation models to their unique business requirements, forcing tradeoffs between generic models with limited performance and complex, expensive customization pipelines that require specialized infrastructure and expertise. Amazon Bedrock removes this complexity by providing a fully managed, secure reinforcement fine-tuning experience. Customers define reward functions using verifiable rule-based graders or AI-based judges, including built-in templates for both objective tasks such as code generation and math reasoning, and subjective tasks such as instruction following or conversational quality. During training, customers can use AWS Lambda functions for custom grading logic, and access intermediate model checkpoints to evaluate, debug, and select the best-performing model, improving iteration speed and training efficiency. All proprietary data remains within AWS’s secure, governed environment throughout the customization process.
Models supported at this launch are: qwen.qwen3-32b and openai.gpt-oss-20b. After fine-tuning completes, customers can immediately use the resulting fine tuned model for on-demand inference through Amazon Bedrock’s OpenAI-compatible APIs - Responses API and Chat Completions API, without any additional deployment steps. To learn more, see the Amazon Bedrock documentation.
Published: 2026-02-17 21:17:00+00:00
Link: https://aws.amazon.com/about-aws/whats-new/2026/02/claude-sonnet-4.6-available-in-amazon-bedrock/
Starting today, Amazon Bedrock supports Claude Sonnet 4.6, which offers frontier performance across coding, agents, and professional work at scale. According to Anthropic, Claude Sonnet 4.6 is their best computer use model yet, allowing organizations to deploy browser-based automation across business tools with near-human reliability. Claude Sonnet 4.6 approaches Opus 4.6 intelligence at a lower cost. It enables faster, high-quality task completion, making it ideal for high-volume coding and knowledge work use cases.
Claude Sonnet 4.6 serves as a direct upgrade to Sonnet 4.5 across use cases that require consistent conversational quality and efficient multi-step orchestration. For search and chat applications, it delivers reliable performance across single and multi-turn exchanges at a price point that makes high-volume deployment practical, maintaining quality standards while optimizing for scale. Developers can leverage Claude Sonnet 4.6’s for agentic workflows, seamlessly filling both lead agent and subagent roles in multi-model pipelines with precise workflow management and context compaction capabilities. Enterprise teams can use Claude Sonnet 4.6 to power domain-specific applications with professional precision, including spreadsheet and financial model creation that accelerates analysis workflows, compliance review processes that require meticulous attention to detail, and data summarization tasks where iteration speed and accuracy are paramount. Claude Sonnet 4.6 requires only minor prompting adjustments from Sonnet 4.5, ensuring smooth migration for existing implementations.
Claude Sonnet 4.6 is now available in Amazon Bedrock. For the full list of available regions, refer to the documentation. To learn more and get started with Claude Sonnet 4.6 in Amazon Bedrock, read the About Amazon blog and visit the Amazon Bedrock console.
Published: 2026-02-17 15:43:00+00:00