As a Cloud Security Architect and DevSecOps Lead, I'm pleased to provide today's summary of key AWS feature updates, focusing on their security implications and current Terraform availability. These enhancements underscore AWS's commitment to robust security, operational efficiency, and developer productivity.
Machine Learning
- SageMaker MLflow now supports Customer-Managed Keys (CMK) with AWS KMS for data encryption. This significant security enhancement allows organizations with stringent compliance requirements to encrypt their MLflow data at rest using their own KMS keys. This provides superior control over encryption, strengthens data governance, and offers comprehensive audit trails via AWS CloudTrail. Architecturally, this means sensitive machine learning artifacts and metadata can now meet enterprise-grade encryption standards, crucial for regulated industries. CMKs must be symmetric and reside in the same AWS account and region as the MLflow App. *Terraform Support*: Supported for
aws_sagemaker_mlflow_tracking_serverwith thekms_key_idargument in Terraform AWS Provider: aws_sagemaker_mlflow_tracking_server in v5.70.0+.
Database
- Amazon Aurora PostgreSQL-Compatible Edition now supports new minor versions (18.4, 17.10, 16.14, 15.18, 14.23). These updates incorporate the latest bug fixes from the PostgreSQL community and Aurora-specific enhancements. From a security and operational perspective, upgrading to the latest minor versions is a critical best practice to address known Common Vulnerabilities and Exposures (CVEs) and ensure optimal database performance and stability. Leveraging automatic minor version upgrades and the AWS Organizations Upgrade Rollout Policy allows for a controlled, phased rollout across environments, minimizing risk to production workloads. *Terraform Support*: Supported for managing Aurora PostgreSQL clusters and specifying
engine_versionvia theaws_rds_clusterresource in Terraform AWS Provider: aws_rds_cluster in v4.5.0+ (new minor versions become available for use with this resource as released).
Contact Center
- Amazon Connect now offers information extraction capabilities. This feature automatically captures key data points from voice and chat interactions, significantly boosting agent and supervisor productivity by reducing manual data entry. Crucially for DevSecOps and compliance, information extraction operates *before* redaction, allowing specific data capture (e.g., account numbers, product names) while still ensuring sensitive values are redacted from recordings and transcripts for privacy and regulatory adherence (e.g., PCI DSS, HIPAA). Extracted data can be used for post-contact work, contact search/review, programmatic access via APIs/Kinesis/S3, and to trigger rule-based actions. *Terraform Support*: While the Amazon Connect instance itself is manageable via
aws_connect_instancein Terraform AWS Provider: aws_connect_instance in v5.20.0+, and related security aspects like KMS encryption for extracted data are supported, direct Terraform support for configuring the specific conversational analytics rules for information extraction is pending provider release (use AWS CLI/SDK or custom provisioner in the interim).