Greetings Cloud Security Architects, DevSecOps Leads, and Terraform Specialists,
Today's AWS update brings a suite of enhancements across various services, focusing on performance, developer experience, and operational efficiency. Crucially, we'll examine the security implications and current Terraform AWS Provider (hashicorp/aws) support for each. Staying abreast of these changes is vital for maintaining robust, compliant, and infrastructure-as-code driven cloud environments.
High Performance Computing
- Amazon DCV: Amazon DCV has significantly expanded its capabilities, now supporting modern Linux distributions like RHEL 10, Rocky Linux 10, CentOS 10, and Ubuntu 26.04, leveraging the Wayland display server. New client updates include a macOS client with gamepad support, and updated Windows and Linux clients. Key enhancements for Windows hosts include URL redirection (links open in local browser) and printer preferences redirection. Performance has also seen improvements with H.264 frame slice support and slice-based multithreaded decoding on Linux and macOS clients, reducing latency. Support for new macOS server versions on Amazon EC2 further broadens deployment options. From a security perspective, Wayland offers a more modern and potentially secure display server architecture. However, features like URL and printer redirection require careful consideration of data exfiltration risks and robust client-side security policies. Always ensure DCV deployments on EC2 leverage secure AMIs, EBS encryption with KMS CMKs, and strong network isolation. *Terraform Support*: The underlying infrastructure for Amazon DCV deployments (EC2 instances, KMS keys, CloudWatch log groups) is fully supported in the Terraform AWS Provider in v4.20.0+ (e.g.,
aws_instance,aws_kms_key,aws_cloudwatch_log_group). The specific DCV software features are generally managed within the EC2 instance configuration or client-side, not directly via AWS API resources.
Containers
- Amazon ECS: Amazon ECS introduces
LiveTaskCount, a new service metric that reports the number of actively running tasks across all desired states (ACTIVATING, RUNNING, DEACTIVATING), including tasks undergoing connection draining. This metric is available at no additional charge and is ideal for precise autoscaling policies and capacity planning. Furthermore, the existingRunningTaskCountmetric in Amazon CloudWatch Container Insights is now also free. This update provides enhanced visibility into the true operational state of ECS services, leading to more accurate autoscaling and improved service resilience. For compliance, accurate metrics are essential for auditing and performance reporting. *Terraform Support*: Theaws_ecs_serviceresource defines ECS services, andaws_appautoscaling_policycan be configured to use these metrics for autoscaling. The availability of these metrics is inherent to the service. Supported in Terraform AWS Provider: aws_ecs_service in v4.0.0+.
Geospatial
- Amazon Location Service: Amazon Location Service now features a comprehensive, interactive API Playground. This tool allows developers to visually explore and test all 14 Places, Routes, and Maps APIs without writing code. Users can make real API calls, see results rendered on an interactive map, adjust parameters, and copy ready-to-use request snippets directly into their applications. While primarily a development tool, the playground helps developers correctly implement API calls, potentially reducing configuration errors that could lead to insecure deployments. Ensure that any API keys or IAM credentials used with the playground adhere to least privilege principles, and maintain strict data privacy for location data. *Terraform Support*: Terraform supports the underlying Amazon Location Service resources (e.g., maps, geofence collections, trackers) via resources like
aws_location_map. The API Playground is a console feature, not a directly provisionable resource. Supported in Terraform AWS Provider: aws_location_map in v3.76.0+.
Machine Learning
- AWS Agent Registry: AWS Agent Registry now supports custom metadata for registry records. Publishers can define a schema once per registry with common attributes, and extend it with type-specific fields (e.g., model family for agents). Fields support free text, URL, Boolean, or choice lists, with validation for required fields. This enables filtering search results by custom metadata. This feature significantly enhances governance and compliance for AI agents by allowing robust data classification, ownership tracking, and audit trails. Defining data classification directly on the agent record helps enforce data handling policies and ensures data integrity. *Terraform Support*: The
aws_bedrock_agent_registryresource is the primary interface. New arguments or blocks within this resource are expected to support custom metadata schema definition and application. Supported in Terraform AWS Provider: aws_bedrock_agent_registry in v5.55.0+.
- Amazon SageMaker: The AWS AI/ML skill for SageMaker has been extended to support several key capabilities: inference optimization, guided end-to-end model customization (fine-tuning), serverful training for traditional ML and generative AI models, MLflow integration for experiment tracking and model lifecycle management, JumpStart model deployment and recommendations, and processing jobs for data preparation. These enhancements provide more granular control and visibility throughout the ML lifecycle. MLflow integration is particularly valuable for auditability and reproducibility, crucial for compliance. Secure data handling, encryption, network isolation, and IAM controls remain paramount across all SageMaker operations. *Terraform Support*: The AWS Provider for Terraform fully supports the underlying Amazon SageMaker services and resources that these capabilities interact with, including
aws_sagemaker_endpoint,aws_sagemaker_model,aws_sagemaker_training_job,aws_sagemaker_processing_job, andaws_sagemaker_notebook_instance. Core SageMaker resources were introduced in v1.10.0+, with continuous updates. Refer to Terraform AWS Provider: SageMaker Service Documentation for comprehensive resource details.
Compute
- AWS Deadline Cloud: AWS Deadline Cloud now improves resource discoverability within its monitor. Users with access to specific queues, fleets, or jobs can directly find these resources, along with their parent IDs and access levels (direct or group-inherited). This simplifies access management and allows teams to share a single monitor entry point regardless of individual permission levels. This feature enhances the principle of least privilege by clarifying access boundaries and improving operational security. Centralized access visibility is beneficial for auditing and compliance, ensuring only authorized personnel interact with specific rendering resources. *Terraform Support*: Terraform manages the underlying Deadline Cloud resources (e.g.,
aws_deadline_queue,aws_deadline_fleet) and the associated IAM policies (aws_iam_policy,aws_iam_role_policy) that define user access. The monitor's display is a UI feature reflecting these underlying IAM permissions. Supported in Terraform AWS Provider: aws_deadline_farm in v5.45.0+.
Networking
- Amazon Route 53: Amazon Route 53 Domains now supports the registration, transfer, and renewal of Premium Domains. These domains automatically appear in domain search and availability results with upfront pricing. Centralizing all domain management, including premium domains, within Route 53 simplifies governance and reduces the attack surface by avoiding multiple registrars. This ensures consistent application of DNSSEC, domain locking, and other security features offered by Route 53. *Terraform Support*: The
aws_route53_domainresource is used for managing domain registrations. The support for premium domains is an extension of this existing resource's functionality. Supported in Terraform AWS Provider: aws_route53_domain in v4.5.0+.
Networking & Content Delivery
- Amazon CloudFront: Amazon CloudFront now allows AWS Lambda Web Functions (Function URLs) to be configured as direct origins for distributions. This enables TLS termination at the edge, caching, custom domains with AWS Certificate Manager (ACM), and crucial security integration with CloudFront Origin Access Control (OAC). OAC restricts direct access to the Lambda Web Function, ensuring all traffic flows securely through CloudFront, providing protection against network-layer DDoS attacks and seamless integration with AWS WAF. Original credentials are preserved for application-level authentication. This is a significant security enhancement for serverless architectures, enforcing a single, secure entry point and aligning with zero-trust principles. *Terraform Support*: This feature requires configuring
aws_cloudfront_distributionto reference anaws_lambda_function_urlas an origin, and utilizingaws_cloudfront_origin_access_controlto secure the connection. New arguments within theoriginblock ofaws_cloudfront_distributionare expected. Supported in Terraform AWS Provider: aws_cloudfront_distribution and aws_cloudfront_origin_access_control in v5.17.0+.
Migration and Transfer
- AWS DataSync: AWS DataSync Enhanced mode now offers expanded support, including Amazon FSx for OpenZFS as a source or destination location, and NFS version 4.2 for on-premises NFS servers and FSx for OpenZFS file systems. A new
ObjectMetadataoption for object storage transfers provides granular control over copying user metadata to the destination. Supporting NFS 4.2 and FSx for OpenZFS expands DataSync's utility while maintaining secure data transfer protocols. TheObjectMetadataoption is critical for compliance and data integrity, allowing control over sensitive metadata and ensuring compatibility with destination systems. Always ensure DataSync agents are deployed securely within private networks and leverage VPC endpoints. *Terraform Support*: Theaws_datasync_location_fsx_openzfsresource would manage FSx for OpenZFS locations, and theaws_datasync_taskresource would likely include new arguments for NFS 4.2 protocol versions and theobject_metadata_optionsblock. Supported in Terraform AWS Provider: aws_datasync_task in v5.50.0+.
Database
- Amazon RDS for SQL Server: Amazon RDS for SQL Server now supports Microsoft SQL Server 2025 CU8 (KB5104822), specifically RDS version 17.00.4075.5.v1. This update includes the latest improvements and critical fixes from Microsoft. Applying the latest Cumulative Updates is a fundamental security best practice, addressing known vulnerabilities and improving database stability. Regular patching is crucial for compliance. Architects should plan for controlled upgrade windows to minimize downtime. *Terraform Support*: The
aws_db_instanceresource is used to manage RDS instances. Theengine_versionargument can be updated to17.00.4075.5.v1to apply this update. Supported in Terraform AWS Provider: aws_db_instance in v5.25.0+.
Object Storage
- Amazon S3 Express One Zone: Amazon S3 Express One Zone, designed for high-performance, low-latency data access, has expanded its availability to four additional AWS Regions: Europe (London), Asia Pacific (Jakarta), Asia Pacific (Hong Kong), and Africa (Cape Town). This brings its total regional availability to 19 regions. Regional expansion enhances data residency and compliance options, allowing customers to keep high-performance data closer to their compute resources while meeting local regulatory requirements. Standard S3 security features like encryption at rest (KMS), IAM policies, and VPC endpoints remain critical for securing data. *Terraform Support*: The
aws_s3_directory_bucketresource is used for S3 Express One Zone. Regional availability is controlled by theproviderblock'sregionargument in Terraform; no new resource arguments are needed for this expansion. Supported in Terraform AWS Provider: aws_s3_directory_bucket in v4.67.0+.
Contact Center
- Amazon Connect: Amazon Connect now enables managers to configure rules for automatically prefilling agent step-by-step guides (After Contact Work - ACW) with information extracted from customer conversations, such as booking identifiers or claim numbers. This significantly reduces manual data entry, improving data accuracy and saving agent time. This feature improves data accuracy and reduces human error, which has security and compliance benefits. However, careful definition of extraction rules is critical to ensure only appropriate, non-sensitive data is automatically populated, or that sensitive data is handled according to strict access controls and redaction policies. Robust logging and auditing of extracted data are essential. *Terraform Support*: This feature likely involves new configurations within Amazon Connect's contact flows or agent experience settings. While
aws_connect_instanceandaws_connect_contact_floware core resources, new arguments or a dedicated resource for ACW guide configuration might be introduced. Supported in Terraform AWS Provider: aws_connect_view in v5.30.0+.
Announcements in this summary
- Amazon DCV High Performance Computing
- Amazon ECS Containers
- Amazon Location Service Geospatial
- AWS Agent Registry Machine Learning
- AWS Deadline Cloud Compute
- Amazon Route 53 Networking
- Amazon SageMaker Machine Learning
- AWS DataSync Migration and Transfer
- Amazon RDS for SQL Server Database
- Amazon CloudFront Networking & Content Delivery
- Amazon S3 Express One Zone Object Storage
- Amazon Connect contact-center