As a Cloud Security Architect, DevSecOps Lead, and Terraform Specialist, I'm pleased to present today's comprehensive summary of new AWS features and service updates. These updates span critical areas from enhanced operational safety and AI innovation to robust data protection and database security, all with a keen eye on their implications for Infrastructure as Code (IaC) and security posture.
Analytics
- Amazon Kinesis Data Streams - DryRun Parameter for API Calls: Amazon Kinesis Data Streams now offers a 'DryRun' optional parameter for
PutRecord,PutRecords,GetRecords,GetShardIterator, andSubscribeToShardAPIs. This feature allows users to validate permissions and request inputs without executing the actual operation, returning aDryRunOperationExceptionupon successful validation. This is a significant enhancement for DevSecOps practices, enabling secure testing of IAM policies and API request validity in pre-production or CI/CD environments, thereby preventing accidental data writes or reads to production streams and bolstering operational safety. Architecturally, it promotes a 'test-before-deploy' mindset for data stream interactions. *Terraform Support*: This is an API-level feature for operational validation and does not directly map to a Terraform resource argument. Terraform manages the *configuration* of Kinesis streams (e.g.,aws_kinesis_stream), but theDryRunparameter is for runtime API calls. Users would leverage AWS CLI/SDK forDryRunoperations. Terraform: aws_kinesis_stream
Customer Engagement
- Amazon Connect Customer - Compact Mode for Analytics Dashboards: Amazon Connect Customer introduces 'compact mode' for analytics dashboards, optimizing screen real estate by reducing widget sizes, fonts, and minimizing filters. This allows supervisors to view more data, such as all team agents, on a single screen without scrolling, facilitating quicker identification of non-adherent agents. While primarily a UI/UX enhancement, it indirectly supports operational efficiency, which can contribute to faster response times for security incidents or compliance checks related to agent activity. *Terraform Support*: This is a console UI feature and does not involve resource configuration managed by Terraform. Terraform: aws_connect_instance
Media Services
- AWS Deadline Cloud - Job Bundle Sharing: AWS Deadline Cloud now enables direct sharing of job bundles (encapsulating job templates, parameters, and asset references) within a queue. These bundles are stored as portable archives in the queue's associated Amazon S3 job attachments bucket, leveraging existing infrastructure. This streamlines collaboration, promotes standardization of job submissions, and reduces manual distribution efforts. From a security perspective, access to shared bundles is governed by existing S3 bucket policies and Deadline Cloud queue access controls, emphasizing the need for robust IAM policies and encryption for data at rest and in transit within S3. *Terraform Support*: Pending provider release. While
aws_deadline_cloud_queueexists, the specific functionality of sharing job bundles within the queue is an operational feature, not a direct resource argument. Terraform would manage the queue and its S3 attachments bucket, but not the content or sharing mechanism of the bundles themselves. Terraform: aws_deadline_cloud_queue
Machine Learning
- Amazon Bedrock - Claude Fable 5.1 Generally Available: Claude Fable 5.1 is now generally available on AWS, offering enhanced reasoning, improved judgment, and reduced confident wrong answers for complex tasks like coding and scientific research. It's designed for high-stakes projects, capable of managing entire software features and code reviews, and can self-identify when it's stuck. Claude Mythos 5.1 (with cyber and bio capabilities) is available with limited access. Anthropic designates Fable 5.1 as a 'Covered Model' with additional data retention and safety policies, complemented by Enterprise Frontier Safeguards (EFS) developed with AWS to maintain data control within customer environments. This is a significant advancement for secure and compliant AI adoption, requiring careful architectural consideration of data governance, IAM roles for model invocation, and adherence to EFS policies. *Terraform Support*: Supported for provisioning resources that utilize this model. For example,
aws_bedrock_provisioned_model_throughputcan be used to provision throughput for specific model ARNs, including Claude Fable 5.1. Supported in Terraform AWS Provider: aws_bedrock_provisioned_model_throughput in v4.59.0+.
Developer Tools
- Amazon Quick - Build Custom Applications with Natural Language: Amazon Quick is now generally available, enabling users to build custom applications by describing them in natural language. It connects directly to various data sources (Salesforce, Jira, databases, etc.) and automatically keeps apps current as underlying data changes, crucially respecting existing identity, authorization, and access control policies. This low-code/no-code platform offers rapid application development while maintaining a strong security posture by integrating with existing access controls. Architects must ensure that the IAM roles assumed by Quick for data source connections adhere strictly to the principle of least privilege, and that data governance policies extend to applications built within Quick. *Terraform Support*: Supported in Terraform AWS Provider: aws_quick_application in v5.55.0+.
Observability
- Amazon CloudWatch Database Insights - Self-Managed PostgreSQL on EC2 Support: Amazon CloudWatch Database Insights now extends its support to self-managed PostgreSQL databases running on Amazon EC2. By utilizing the CloudWatch agent, it collects health and performance data, providing a unified fleet view with live performance metrics, wait event analysis, query-level statistics, and host metrics, mirroring the experience for RDS and Aurora. This significantly enhances observability for critical self-managed database workloads, crucial for identifying performance bottlenecks and potential security anomalies. Security considerations include securing the CloudWatch agent installation with least privilege IAM roles and ensuring encrypted data transmission to CloudWatch. This aids in compliance by providing comprehensive monitoring and audit capabilities. *Terraform Support*: Supported for configuring CloudWatch resources to consume these insights. While the CloudWatch agent installation on EC2 is typically managed via user data, SSM, or configuration management tools, the related CloudWatch dashboards, alarms, and log groups can be managed using Terraform. For example,
aws_cloudwatch_dashboardoraws_cloudwatch_log_group. Supported in Terraform AWS Provider in v5.55.0+.
Data Protection
- AWS Backup - Backup and Restore for More Than 1,000 S3 Buckets: AWS Backup now supports backing up and restoring over 1,000 Amazon S3 buckets per account, aligning with the Amazon S3 bucket quota. This removes a previous limitation, allowing comprehensive data protection for large-scale S3 deployments. This is a critical scalability enhancement for data protection strategies, ensuring that all general-purpose S3 buckets can be centrally managed for backup and recovery. Security architects must verify that AWS Backup's IAM roles have appropriate permissions to access the expanded number of S3 buckets, and that backup plans meet RPO/RTO and compliance requirements across the entire S3 estate. *Terraform Support*: Supported for defining backup plans and selections that can now encompass more than 1,000 S3 buckets. The underlying capacity increase is a service-level enhancement. You can manage
aws_backup_planandaws_backup_selectionto include S3 resources. Supported in Terraform AWS Provider: aws_backup_vault (and relatedaws_backup_plan,aws_backup_selectionresources) in v4.20.0+.
Database
- Amazon RDS Custom for SQL Server - Latest CU+GDR Support: Amazon RDS Custom for SQL Server now supports SQL Server 2019 CU32+GDR (KB5102335, RDS version 15.00.4480.2.v1) and SQL Server 2022 CU25+GDR (KB5101347, RDS version 16.00.4262.2.v1). These updates specifically patch critical vulnerabilities (CVE-2026-47295, CVE-2026-47296, CVE-2026-54118, and CVE-2026-55002). This is a vital security update, enabling customers to mitigate known vulnerabilities and maintain a strong security posture for their self-managed SQL Server instances within the RDS Custom framework. DevSecOps teams should prioritize applying these patches as part of their regular patching cycles to ensure compliance and reduce the attack surface. *Terraform Support*: Supported by specifying the updated
engine_versionforaws_rds_custom_instance. For example,engine_version = "15.00.4480.2.v1". Supported in Terraform AWS Provider: aws_rds_custom_instance in v3.37.0+.