AWS announces support for collecting systemd journal (journald) logs with the Amazon CloudWatch agent. You can now configure the CloudWatch agent to read log entries directly from the systemd journal on Linux instances and send them to Amazon CloudWatch Logs, without first writing those logs to files on disk.
Many modern Linux distributions, including Amazon Linux 2023, use systemd journal as the primary logging system and no longer write traditional text log files such as /var/log/messages by default. Previously, collecting these logs with the CloudWatch agent required additional configuration to export the journal to files on disk. With this launch, the CloudWatch agent reads journald entries natively, preserving the structured metadata that journald captures, such as the systemd unit, priority, and process information. You can filter log entries using systemd units, journal priority levels, and journal field matches, and you can apply regular expression filters before logs are published to CloudWatch Logs. This helps you reduce noise and control log volume and costs.
Support for journald in the CloudWatch agent is available in all AWS Commercial Regions and GovCloud(US) regions. Standard Amazon CloudWatch Logs pricing applies for ingested logs. For more information, see Amazon CloudWatch Pricing page.
AWS announces support for collecting systemd journal (journald) logs with the Amazon CloudWatch agent. You can now configure the CloudWatch agent to read log entries directly from the systemd journal on Linux instances and send them to Amazon CloudWatch Logs, without first writing those logs to files on disk.
Many modern Linux distributions, including Amazon Linux 2023, use systemd journal as the primary logging system and no longer write traditional text log files such as /var/log/messages by default. Previously, collecting these logs with the CloudWatch agent required additional configuration to export the journal to files on disk. With this launch, the CloudWatch agent reads journald entries natively, preserving the structured metadata that journald captures, such as the systemd unit, priority, and process information. You can filter log entries using systemd units, journal priority levels, and journal field matches, and you can apply regular expression filters before logs are published to CloudWatch Logs. This helps you reduce noise and control log volume and costs.
Support for journald in the CloudWatch agent is available in all AWS Commercial Regions and GovCloud(US) regions. Standard Amazon CloudWatch Logs pricing applies for ingested logs. For more information, see Amazon CloudWatch Pricing page.
To get started, update to the latest version of the CloudWatch agent and add a journald section to your agent configuration file. To learn more, see Manually create or edit the CloudWatch agent configuration file in the Amazon CloudWatch User Guide.
Starting today, Amazon Aurora MySQL-Compatible Edition 3 (compatible with MySQL 8.0) supports MySQL 8.0.45, which includes community MySQL fixes and Aurora-specific improvements. For detailed information on this release, refer to Aurora MySQL 3 and MySQL 8.0.45 release notes.
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 clusters in phases. You can perform minor version upgrades in-place or via snapshot restore. This release is supported in all AWS Regions where Aurora MySQL is available.
Amazon Aurora is designed for high performance and availability at global scale with full MySQL compatibility. It provides scale-to-zero serverless compute, Aurora Global Database for multi-Region resilience, Aurora I/O-Optimized for improved price performance on I/O-intensive workloads, and built-in security and continuous backups. To get started, take a look at Aurora’s getting started page.
Starting today, Amazon Aurora MySQL-Compatible Edition 3 (compatible with MySQL 8.0) supports MySQL 8.0.45, which includes community MySQL fixes and Aurora-specific improvements. For detailed information on this release, refer to Aurora MySQL 3 and MySQL 8.0.45 release notes. 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 clusters in phases. You can perform minor version upgrades in-place or via snapshot restore. This release is supported in all AWS Regions where Aurora MySQL is available. Amazon Aurora is designed for high performance and availability at global scale with full MySQL compatibility. It provides scale-to-zero serverless compute, Aurora Global Database for multi-Region resilience, Aurora I/O-Optimized for improved price performance on I/O-intensive workloads, and built-in security and continuous backups. To get started, take a look at Aurora’s getting started page.
Por qué la incorporación es importante para la experiencia de los agentes y cómo hacerlo bien.
Por: Pooja Dhaka y Garima Sikdar
Los usuarios no forman una primera impresión solo por la calidad del resultado. Su experiencia comienza en el momento en que abren el agente
Para cualquier equipo que construya un agente, factores como la precisión, la latencia, las alucinaciones y el manejo del contexto son lo más importante. Y con razón: moldean cada interacción. Pero los usuarios no forman una primera impresión solo por la calidad del resultado. Su experiencia comienza en el momento en que abren el agente, y es importante que tratemos las interacciones iniciales con tanto cuidado como las que vienen después.
Para entender qué causa una buena primera impresión, nuestra investigación de experiencia de usuario se centró en esos minutos cruciales de apertura con usuarios que prueban un agente de ventas por primera vez.
El agente estaba pensado y construido para gerentes de ventas y representantes que querían escalar el alcance personalizado. Antes de que los usuarios pudieran ver cualquier resultado, tenían que completar la incorporación: proporcionar fuentes de conocimiento, configurar al agente, ejecutar simulaciones… todo a través de una interfaz basada en chat.
Cuando observamos a los usuarios probar el agente por primera vez, descubrimos que su impresión se veía moldeada de manera significativa por cómo la experiencia de incorporación enmarcaba el valor, la capacidad y el papel. Basándonos en lo que aprendimos, llegamos a cinco principios para diseñar esos primeros momentos.
01
Primeras impresiones:
Entregar el «momento ajá» desde el principio
Al incorporar leads con la configuración – conectar cuentas, establecer preferencias, describir su rol – se pide a los usuarios que inviertan antes de haber visto lo que el agente puede hacer. En nuestro estudio, los responsables de ventas que tuvieron que proporcionar múltiples fuentes de conocimiento antes de ver el resultado de cualquier agente, encontraron dificultades para mantener la motivación durante la configuración.
La solución no es eliminar la configuración. Es reordenarla. Muestren primero un resultado útil y luego pidan contexto cuando el usuario entienda por qué importa. Si el agente necesita fuentes de conocimiento, generar un borrador base a partir de una entrada mínima y luego mostrar cómo añadir fuentes mejora la calidad.
La experiencia aporta valor primero, lo que demuestra lo que el agente puede hacer antes de solicitar contexto adicional para mejorar el resultado.
Dividir la configuración en pequeñas acciones con recompensas inmediatas mantiene a los usuarios en marcha.
02
La secuenciación importa:
Ofrecer pequeños pasos con recompensas inmediatas
El principio 1 aborda cuándo los usuarios ven valor durante la configuración. Pero también consideremos cómo llegan ahí. ¿Cuál es la estructura y el ritmo más efectivos para cada paso de configuración? En nuestro estudio, el agente de ventas requería diferentes fuentes de conocimiento y datos de prospectos antes de poder simular el contacto mediante la redacción de correos electrónicos. Pedir a los responsables que completaran todo eso hacía que la incorporación resultara abrumadora.
Dividir la configuración en pequeñas acciones con recompensas inmediatas mantiene el avance de los usuarios. Diseñar la completación parcial, donde el agente produce una salida útil incluso con entradas faltantes, significa que los usuarios no se quedan bloqueados por una configuración incompleta. Por defecto «guardar y reanudar» para flujos más largos, y mostrar qué se ha hecho y qué viene después, mantiene el impulso informativo en lugar de performativo. Permitan a los usuarios ver qué puede hacer el agente con lo que tiene, y luego muestren qué se podría desbloquear con entradas adicionales.
03
Impulso visible:
Dar señales claras de que el agente entiende y funciona
Cuando un agente tarda en procesar, la ausencia de retroalimentación genera sus propios problemas. Observamos que los participantes que realizaban simulaciones del contacto con el agente de ventas no tenían una señal clara de lo que el agente estaba haciendo durante el procesamiento. La falta de señales de progreso llevó a repetidos prompts y a incertidumbre sobre si el sistema había entendido la solicitud.
Las señales de progreso no necesitan ser detalladas. Incluso descripciones en lenguaje sencillo de lo que hace el agente («analizar datos de prospectos», «redactar la divulgación», «ejecutar simulaciones») ayudan mucho a mantener a los usuarios orientados. Cuando los usuarios pueden ver que el agente entiende su solicitud y trabaja de manera activa en ella, es menos probable que lo intenten de nuevo o duden de ella. También ayuda establecer expectativas de tiempo aproximadas cuando sea posible y dejar claros los estados finales para que los usuarios sepan si algo está terminado, completado de manera parcial, necesita su aportación o ha fallado.
La personalización efectiva se produce a través de preferencias aprendidas, contexto inferido y valores predeterminados inteligentes, no a largas conversaciones de incorporación.
04
Personalización sin fricciones:
Empezar con una salida sencilla que los usuarios puedan refinar más adelante
La salida del agente que refleja la audiencia, las restricciones y las preferencias del usuario siempre resultará más útil que algo genérico. Pero cuanto más contexto pida un agente desde el principio, más pesada se siente la configuración. Aunque los responsables de ventas esperaban que el agente adaptara el contacto a sus prospectos y estilo de venta, resistieron de manera comprensible los largos flujos de configuración para lograrlo.
La personalización efectiva se produce a través de preferencias aprendidas, contexto inferido y valores predeterminados inteligentes, no en largas conversaciones de incorporación. Empezar con valores predeterminados sensatos permite a los usuarios refinar tras ver la salida inicial en lugar de configurar antes de haber visto nada. Pedir contexto solo cuando cambiará el resultado de manera material, retenerlo una vez proporcionado y mostrar el impacto mediante comparaciones antes y después (por ejemplo, un correo de divulgación con y sin detalles específicos de cada prospecto), hace que el valor de compartir información sea concreto en lugar de abstracto.
Al dar a conocer las sugerencias relevantes desde el principio, el agente reduce la fricción y fomenta la participación antes de requerir contexto adicional.
05
Establezcan expectativas desde el principio:
Mostrar a los usuarios dónde el agente puede ofrecerles más valor
Los usuarios están acostumbrados a herramientas como Copilot y ChatGPT, y esas experiencias moldean lo que prueban primero. En nuestro estudio, los representantes de ventas asumían que el agente podía extraer información de la web y responder preguntas generales. Pero estaba diseñado para redactar y enviar correos de ventas de manera autónoma, y era necesario la incorporación para llegar a ello.
Mostrar algunas tareas de ejemplo al principio ayuda a los usuarios a entender para qué está diseñado el agente. Mantener el lenguaje simple y orientado a la acción facilita el intento. Tras cada paso, sugerir qué hacer a continuación mantiene a los usuarios en movimiento sin necesidad de una guía completa.
El efecto de composición
Cada uno de estos principios se basa en el otro de manera mutua. Cuando los usuarios ven valor desde el principio, están más dispuestos a invertir. Cuando el progreso es visible, no cuestionan el sistema ni se desconectan. Cuando la personalización crece con el tiempo en lugar de desde el principio, empezar se siente más ligero. Las capacidades de superficie ayudan a los usuarios a conectar con lo que el agente en verdad está hecho para hacer. Y cuando la configuración se divide en pasos más pequeños, los usuarios siguen su avance en lugar de abandonar. Juntos, estos factores moldean si los primeros cinco minutos generan confianza o la erosionan.
Aunque estos principios surgieron de la investigación de incorporación, se extienden a cualquier momento en que los usuarios interactúen por primera vez con un agente o cuando su comportamiento cambia. Cada nueva capacidad, flujo de trabajo o cambio de comportamiento es otra primera impresión. El objetivo no es tan solo explicar cómo funciona el agente, sino ayudar a los usuarios a ganar confianza, sentirse en control y tener confianza para dar el siguiente paso.
NVIDIA’s Cosmos3-Edge, Cosmos3-Nano, and Cosmos3-Super models are now available on Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These three models form the Cosmos 3 family of open, frontier omnimodal world models for physical AI, enabling customers to build robots, autonomous vehicles, and vision AI that perceive, reason, plan, and act in the physical world.
These models address different physical AI challenges with specialized capabilities:
Cosmos3-Edge is engineered for on-device robot control and real-time visual reasoning on edge hardware. This 4B-parameter omni-model (with a 2B Nemotron-based reasoner) operates at robot-control resolution (640×360), delivering real-time reasoning and generating 32 actions per inference at 15 Hz on NVIDIA Jetson Thor. It supports 256p and 480p video at 12–30 FPS, bringing frontier physical AI capabilities directly to embedded systems.
Cosmos3-Nano excels in physics-aware world generation and physical reasoning as a compact 16B-parameter omnimodal model. It processes combinations of text, image, video, audio, and action trajectories to produce corresponding outputs, enabling robots and vision AI agents to reason using prior knowledge, physics understanding, and common sense. It supports chain-of-thought reasoning over text, images, and video with resolutions up to 720p.
Cosmos3-Super provides the highest-fidelity world generation and simulation in the Cosmos 3 family at 64B parameters. It jointly processes and generates language, images, video, audio, and action sequences within a unified Mixture-of-Transformers architecture, supporting resolutions up to 720p across multiple aspect ratios. Ideal for large-scale simulation, synthetic data generation, and policy learning workflows.
With SageMaker JumpStart, customers can deploy any of these models with just a few clicks to address their specific AI use cases.
To get started with these models, navigate to the SageMaker JumpStart model catalog in the SageMaker console or use the SageMaker Python SDK to deploy the models to your AWS account. For more information about deploying and using foundation models in SageMaker JumpStart, see the Amazon SageMaker JumpStart documentation.
NVIDIA’s Cosmos3-Edge, Cosmos3-Nano, and Cosmos3-Super models are now available on Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These three models form the Cosmos 3 family of open, frontier omnimodal world models for physical AI, enabling customers to build robots, autonomous vehicles, and vision AI that perceive, reason, plan, and act in the physical world.
These models address different physical AI challenges with specialized capabilities:
Cosmos3-Edge is engineered for on-device robot control and real-time visual reasoning on edge hardware. This 4B-parameter omni-model (with a 2B Nemotron-based reasoner) operates at robot-control resolution (640×360), delivering real-time reasoning and generating 32 actions per inference at 15 Hz on NVIDIA Jetson Thor. It supports 256p and 480p video at 12–30 FPS, bringing frontier physical AI capabilities directly to embedded systems.
Cosmos3-Nano excels in physics-aware world generation and physical reasoning as a compact 16B-parameter omnimodal model. It processes combinations of text, image, video, audio, and action trajectories to produce corresponding outputs, enabling robots and vision AI agents to reason using prior knowledge, physics understanding, and common sense. It supports chain-of-thought reasoning over text, images, and video with resolutions up to 720p.
Cosmos3-Super provides the highest-fidelity world generation and simulation in the Cosmos 3 family at 64B parameters. It jointly processes and generates language, images, video, audio, and action sequences within a unified Mixture-of-Transformers architecture, supporting resolutions up to 720p across multiple aspect ratios. Ideal for large-scale simulation, synthetic data generation, and policy learning workflows.
With SageMaker JumpStart, customers can deploy any of these models with just a few clicks to address their specific AI use cases.
To get started with these models, navigate to the SageMaker JumpStart model catalog in the SageMaker console or use the SageMaker Python SDK to deploy the models to your AWS account. For more information about deploying and using foundation models in SageMaker JumpStart, see the Amazon SageMaker JumpStart documentation.
Meta’s Muse-Glimmer-30B and Alibaba’s Qwen 3.8-27B models are now available on Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These two models bring specialized capabilities spanning autonomous local agentic workflows and multimodal long-horizon reasoning, enabling customers to deploy high-performance, scalable AI solutions on AWS infrastructure.
These models address different enterprise AI challenges with specialized capabilities:
Muse-Glimmer-30B is engineered for autonomous agentic tasks with multi-step reasoning, tool use, and failure recovery. This 30B-parameter dense model from Meta Superintelligence Lab combines a dedicated ~1.8B ViT-G/14 perception encoder with interleaved text and image inputs, a 131K+ context window, and selectable reasoning strength (low through extra-high). Released under Apache 2.0, it handles sequential tool calls, recovers from failures, and operates entirely without cloud infrastructure which is ideal for always-on enterprise agents.
Qwen 3.8-27B excels in coding, multi-step agentic tasks, and multimodal understanding across text, images, and video. A dense 27B-parameter native vision-language model with a 262K context window (extendable to ~1M via YaRN scaling), it delivers substantial gains over its predecessor with adjustable reasoning effort levels. Scoring 61.7 on SWE-bench Pro and running at ~17GB quantized, it carries complex multi-step tasks through to completion with greater reliability.
With SageMaker JumpStart, customers can deploy any of these models with just a few clicks to address their specific AI use cases.
To get started with these models, navigate to the SageMaker JumpStart model catalog in the SageMaker console or use the SageMaker Python SDK to deploy the models to your AWS account. For more information about deploying and using foundation models in SageMaker JumpStart, see the Amazon SageMaker JumpStart documentation.
Meta’s Muse-Glimmer-30B and Alibaba’s Qwen 3.8-27B models are now available on Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These two models bring specialized capabilities spanning autonomous local agentic workflows and multimodal long-horizon reasoning, enabling customers to deploy high-performance, scalable AI solutions on AWS infrastructure.
These models address different enterprise AI challenges with specialized capabilities:
Muse-Glimmer-30B is engineered for autonomous agentic tasks with multi-step reasoning, tool use, and failure recovery. This 30B-parameter dense model from Meta Superintelligence Lab combines a dedicated ~1.8B ViT-G/14 perception encoder with interleaved text and image inputs, a 131K+ context window, and selectable reasoning strength (low through extra-high). Released under Apache 2.0, it handles sequential tool calls, recovers from failures, and operates entirely without cloud infrastructure which is ideal for always-on enterprise agents.
Qwen 3.8-27B excels in coding, multi-step agentic tasks, and multimodal understanding across text, images, and video. A dense 27B-parameter native vision-language model with a 262K context window (extendable to ~1M via YaRN scaling), it delivers substantial gains over its predecessor with adjustable reasoning effort levels. Scoring 61.7 on SWE-bench Pro and running at ~17GB quantized, it carries complex multi-step tasks through to completion with greater reliability.
With SageMaker JumpStart, customers can deploy any of these models with just a few clicks to address their specific AI use cases.
To get started with these models, navigate to the SageMaker JumpStart model catalog in the SageMaker console or use the SageMaker Python SDK to deploy the models to your AWS account. For more information about deploying and using foundation models in SageMaker JumpStart, see the Amazon SageMaker JumpStart documentation.
Amazon Redshift now supports Amazon Kinesis Data Streams (KDS) record sizes up to 10 MiB—a 10x increase from the previous 1 MiB limit—fully matching the expanded maximum record size in Amazon KDS. This means you can stream significantly larger payloads directly into Amazon Redshift without splitting records, simplifying your ingestion pipelines and unlocking new use cases for high-volume, large-record workloads.
Amazon Redshift support for 10MiB record size in Amazon KDS streams is now available in all commercial AWS regions where Amazon Redshift is available. For more information on direct streaming ingestion into Amazon Redshift, see the Amazon Redshift streaming documentation. For more information on 10MiB record support in Amazon KDS, see the Amazon KDS documentation.
Amazon Redshift now supports Amazon Kinesis Data Streams (KDS) record sizes up to 10 MiB—a 10x increase from the previous 1 MiB limit—fully matching the expanded maximum record size in Amazon KDS. This means you can stream significantly larger payloads directly into Amazon Redshift without splitting records, simplifying your ingestion pipelines and unlocking new use cases for high-volume, large-record workloads.
Amazon Redshift support for 10MiB record size in Amazon KDS streams is now available in all commercial AWS regions where Amazon Redshift is available. For more information on direct streaming ingestion into Amazon Redshift, see the Amazon Redshift streaming documentation. For more information on 10MiB record support in Amazon KDS, see the Amazon KDS documentation.
Amazon Bedrock AgentCore is now available in two additional AWS Regions: US West (N. California) and Asia Pacific (Hyderabad). Amazon Bedrock AgentCore is the platform to build, connect, and optimize agents. It helps engineers ship agents fast with any framework and any model, connect them to enterprise systems and tools, and optimize them continuously, with security enforced at the infrastructure layer that agents can’t bypass.
With this expansion, customers in these regions can build and run agents closer to their end users with lower latency. AgentCore capabilities including agent runtime, identity and access control, policy management, session persistence, tool connectivity, evaluations and observability are available in these regions at launch.
Amazon Bedrock AgentCore is now available in two additional AWS Regions: US West (N. California) and Asia Pacific (Hyderabad). Amazon Bedrock AgentCore is the platform to build, connect, and optimize agents. It helps engineers ship agents fast with any framework and any model, connect them to enterprise systems and tools, and optimize them continuously, with security enforced at the infrastructure layer that agents can’t bypass.
With this expansion, customers in these regions can build and run agents closer to their end users with lower latency. AgentCore capabilities including agent runtime, identity and access control, policy management, session persistence, tool connectivity, evaluations and observability are available in these regions at launch.
For more information on AgentCore, visit the AgentCore product page or the AgentCore Developer Guide. To learn about pricing, visit AgentCore pricing. For region availability, visit Supported AWS Regions.
Amazon Connect Customer now supports generative AI-powered summaries, real-time call analytics, and real-time rules in the Africa (Cape Town) Region. These capabilities extend the post-contact conversational analytics already available in the Region, giving companies access to AI-powered insights that help them improve customer experience, agent performance, and operational efficiency.
With generative AI-powered summaries, supervisors and human-agents receive concise, AI-generated summaries of customer interactions immediately after a contact ends, eliminating the need for manual note-taking and reducing after-contact work. Real-time call analytics provide live transcription enabling supervisors to monitor interactions as they happen and intervene when needed to improve outcomes. Real-time rules automatically categorize contacts and alert supervisors based on keywords, sentiment, and other criteria detected during a live interaction, enabling immediate action such as sending notifications or generating tasks.
Amazon Connect Customer now supports generative AI-powered summaries, real-time call analytics, and real-time rules in the Africa (Cape Town) Region. These capabilities extend the post-contact conversational analytics already available in the Region, giving companies access to AI-powered insights that help them improve customer experience, agent performance, and operational efficiency.
With generative AI-powered summaries, supervisors and human-agents receive concise, AI-generated summaries of customer interactions immediately after a contact ends, eliminating the need for manual note-taking and reducing after-contact work. Real-time call analytics provide live transcription enabling supervisors to monitor interactions as they happen and intervene when needed to improve outcomes. Real-time rules automatically categorize contacts and alert supervisors based on keywords, sentiment, and other criteria detected during a live interaction, enabling immediate action such as sending notifications or generating tasks.
To learn more about Connect Customer conversational analytics capabilities, refer to the Amazon Connect Customer Administrator Guide or visit the Amazon Connect Customer website. For a complete list of conversational analytics features available by Region, refer to Availability of Connect Customer features by Region.
Amazon Connect Customer now automatically refreshes schedule metrics on the scheduling page, giving managers immediate visibility into the impact of schedule updates. For example, when a new team meeting is added for 10 agents from 10 AM – 11 AM, metrics such as net available headcount and projected service level are automatically updated. This gives workforce managers up-to-date visibility into how schedule changes affect staffing coverage, so they can make faster, more informed decisions throughout the day.
This feature is available in all AWS Regions where Amazon Connect Customer agent scheduling is available. To learn more about Amazon Connect Customer agent scheduling, click here .
Amazon Connect Customer now automatically refreshes schedule metrics on the scheduling page, giving managers immediate visibility into the impact of schedule updates. For example, when a new team meeting is added for 10 agents from 10 AM – 11 AM, metrics such as net available headcount and projected service level are automatically updated. This gives workforce managers up-to-date visibility into how schedule changes affect staffing coverage, so they can make faster, more informed decisions throughout the day.
This feature is available in all AWS Regions where Amazon Connect Customer agent scheduling is available. To learn more about Amazon Connect Customer agent scheduling, click here .
Amazon Redshift now integrates with the Agent Toolkit for AWS, enabling you to build, query, troubleshoot, and migrate to Amazon Redshift data warehouses and date lakes directly from AI agents such as Claude Code, Kiro, and Cursor. The integration pairs the AWS MCP (Model Context Protocol) server, which provides authenticated AWS API execution on your behalf — with Redshift skills: curated packages of tested procedures and reference material that help AI agents complete Redshift tasks more effectively.
The Redshift skills cover SQL syntax references to reduce query generation errors, metadata discovery to explore schemas and data without writing SQL by hand, data loading patterns, materialized view best practices, function and data type guidance, and extensions such as Qualify, Pivot, and Super. It also guides end-to-end data warehouse migrations to Amazon Redshift, including discovery, schema and SQL conversion, data movement, validation, and performance comparison. We will continue to expand these skills with additional capabilities over time.
The skills work with provisioned clusters and Serverless workgroups, require no changes to existing infrastructure, and are available at no additional charge in all AWS Regions where Amazon Redshift and the AWS MCP Server are offered.
To get started, install the aws-data-analytics plugin in your agent, which bundles the MCP Server configuration and Redshift skills in a single step. Agents with MCP Server access can also discover and load skills at runtime without pre-installation. For setup instructions, see the Agent Toolkit documentation or the Amazon Redshift skills documentation.
Amazon Redshift now integrates with the Agent Toolkit for AWS, enabling you to build, query, troubleshoot, and migrate to Amazon Redshift data warehouses and date lakes directly from AI agents such as Claude Code, Kiro, and Cursor. The integration pairs the AWS MCP (Model Context Protocol) server, which provides authenticated AWS API execution on your behalf — with Redshift skills: curated packages of tested procedures and reference material that help AI agents complete Redshift tasks more effectively.
The Redshift skills cover SQL syntax references to reduce query generation errors, metadata discovery to explore schemas and data without writing SQL by hand, data loading patterns, materialized view best practices, function and data type guidance, and extensions such as Qualify, Pivot, and Super. It also guides end-to-end data warehouse migrations to Amazon Redshift, including discovery, schema and SQL conversion, data movement, validation, and performance comparison. We will continue to expand these skills with additional capabilities over time.
The skills work with provisioned clusters and Serverless workgroups, require no changes to existing infrastructure, and are available at no additional charge in all AWS Regions where Amazon Redshift and the AWS MCP Server are offered.
To get started, install the aws-data-analytics plugin in your agent, which bundles the MCP Server configuration and Redshift skills in a single step. Agents with MCP Server access can also discover and load skills at runtime without pre-installation. For setup instructions, see the Agent Toolkit documentation or the Amazon Redshift skills documentation.
