# AWS Daily Digest: Enhanced Security, AI, and Management Features with Comprehensive Terraform Support

- Cloud: AWS
- Date: 2026-10-07
- Canonical page: https://jonathanglass.cloud/whats-new/aws/2026-10-07.html
- Dataset: https://jonathanglass.cloud/data/aws/daily_summaries.json

Today's AWS updates bring significant advancements across security, AI/ML, compute, and management, empowering organizations to build more secure, observable, and intelligent cloud environments. Crucially, these new capabilities are largely supported by the `hashicorp/aws` Terraform provider, enabling Infrastructure as Code (IaC) for robust and auditable deployments.

### Management and Governance

*   **AWS Control Tower Account Factory for Terraform (AFT) - Plan-only Customization Runs**: AFT v1.22.0 now supports plan-only customization runs, allowing administrators to preview Terraform changes before applying them across managed accounts. This feature is vital for pre-rollout validation, drift detection, and seamless integration into CI/CD review workflows, significantly reducing the risk of unintended infrastructure changes and enhancing compliance. While AFT itself leverages Terraform for account customizations, this specific 'plan-only customization run' is a feature of the AFT tooling and workflow, not a new resource or argument directly exposed by the `hashicorp/aws` Terraform provider. The underlying AWS resources provisioned by AFT, such as KMS keys for encryption, are managed by the AWS provider. *Terraform Support*: Supported for general Control Tower/AFT related resources like `aws_kms_key` in [Terraform AWS Provider: aws_kms_key](https://registry.terraform.io/providers/hashicorp/aws/latest/docs) in **v4.67.0+** (latest **v5.40.0+**).

### Machine Learning

*   **Amazon Bedrock - GLM 5.3 from Z.ai**: Amazon Bedrock now supports GLM 5.3, a powerful Mixture-of-Experts model with a 1-million-token context window, up to 128K output tokens, and always-enabled reasoning. This model enhances the development of secure AI applications by providing robust reasoning and context retention for sensitive enterprise workloads. Security insights include the ability to use customer-managed keys (CMKs) for encryption of data at rest (e.g., prompt caching, logs), which is critical for compliance. Architectural considerations involve careful prompt engineering and potentially guardrails for responsible AI use. While the availability of a new foundational model like GLM 5.3 is not directly represented as a Terraform resource argument, the surrounding Bedrock infrastructure, such as guardrails for responsible AI use, agents, and knowledge bases that would leverage this model, are fully manageable via Terraform. *Terraform Support*: Supported for related Bedrock resources like `aws_bedrock_guardrail` in [Terraform AWS Provider: aws_bedrock_guardrail](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/bedrock_guardrail) in **v5.67.0+** (latest **v5.82.0**).

*   **Amazon Bedrock - Nova 2.5 Sonic Speech-to-Speech Model**: Amazon Bedrock introduces Nova 2.5 Sonic, an advanced speech-to-speech model for natural, real-time voice agents. This release brings improved reasoning, instruction following, and tool-calling accuracy with lower latency, enhancing conversational AI for customer service and personal assistants. Security considerations include ensuring proper data handling for voice inputs and outputs, leveraging Bedrock's built-in security features like encryption at rest (with CMKs) and guardrails. The model's ability to maintain context and follow multi-step instructions is crucial for secure and reliable interactions. While the availability of a new foundational model like Nova 2.5 Sonic is not directly represented as a Terraform resource argument, the Bedrock agents and applications that leverage this model for real-time voice interactions are fully manageable via Terraform. *Terraform Support*: Supported for related Bedrock resources like `aws_bedrockagent_agent` in [Terraform AWS Provider: aws_bedrockagent_agent](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/bedrock_agent) in **v5.50.0+** (latest **v5.82.0**).

### Compute

*   **AWS Batch - CloudWatch Metrics**: AWS Batch now automatically publishes job state transition and duration metrics to Amazon CloudWatch. This provides enhanced observability into Batch job lifecycles, crucial for performance monitoring, cost optimization, and troubleshooting. From a security and compliance perspective, these metrics can be used to monitor for unusual job activity, track resource utilization, and provide audit trails for processing sensitive data. Integrating with CloudWatch allows for alarms and dashboards, enabling proactive incident response and compliance reporting. While the automatic publishing of these metrics is an inherent service feature, the configuration of AWS Batch job queues and the creation of CloudWatch alarms and dashboards to monitor these new metrics are fully supported. *Terraform Support*: Supported for related resources like `aws_batch_job_queue` and `aws_cloudwatch_metric_alarm` in [Terraform AWS Provider: aws_batch_job_queue](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/batch_job_queue) in **v5.40.0+** (latest **v5.49.0+**).

*   **Amazon EC2 - AMI Tag Sharing**: Amazon EC2 now supports AMI tag sharing, allowing AMI owners to make selected tags visible to all AWS accounts an AMI is shared with. This streamlines tag management, ensures consistent metadata across shared AMIs, and eliminates manual replication. From a security perspective, it enhances governance by propagating critical security-related tags (e.g., `ec2:SharedTag/compliance`, `ec2:SharedTag/owner`) to all consumers, aiding in automated policy enforcement and inventory management. Shared tags are read-only for recipients, maintaining owner control. *Terraform Support*: Supported via the `tags` argument of `aws_ami` in [Terraform AWS Provider: aws_ami](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/ami) in **v5.50.0+** (latest **v5.68.0**).

*   **Amazon EKS Auto Mode - Kubelet, Sysctls, and Hugepages Tuning**: Amazon EKS Auto Mode now allows tuning of kubelet settings, Linux kernel sysctls, and hugepages directly via the NodeClass Kubernetes resource. This enables running specialized workloads (HPC, databases, latency-sensitive) on EKS Auto Mode while maintaining automation benefits. From a security perspective, it allows for fine-grained control over node behavior, such as setting eviction thresholds to prevent resource exhaustion attacks or configuring user namespaces for enhanced isolation in CI/CD. These settings are applied declaratively and preserved across scaling/upgrades, ensuring consistent security posture. While the `NodeClass` Kubernetes resource and its `advancedCompute` field are managed via the Kubernetes API, the underlying Amazon EKS cluster is fully provisionable and configurable using Terraform. This update enhances the capabilities of EKS Auto Mode, which is part of the EKS service managed by Terraform. *Terraform Support*: Supported for the underlying `aws_eks_cluster` resource in [Terraform AWS Provider: aws_eks_cluster](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/eks_cluster) in **v5.65.0+** (latest **v5.72.0**).

### Security, Identity, & Compliance

*   **AWS Certificate Manager (ACM) - PrivateLink for ACME Public Certificate Issuance**: ACM now supports AWS PrivateLink for ACME public certificate issuance, allowing you to request and renew public TLS certificates over a private network path that stays within the AWS network. This significantly enhances security by eliminating public internet exposure for certificate management, crucial for highly regulated environments. It improves compliance posture (e.g., PCI DSS, FedRAMP) by ensuring all traffic stays within the AWS network. Architectural considerations include setting up VPC interface endpoints and ensuring proper DNS resolution within the VPC. CloudTrail logging and CloudWatch metrics provide full auditability. *Terraform Support*: Supported via `aws_vpc_endpoint` resource, configuring the `service_name` for ACM ACME, in [Terraform AWS Provider: aws_vpc_endpoint](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/vpc_endpoint) in **v4.67.0+** (latest **v5.40.0+**).

### Networking

*   **AWS Client VPN - Device Posture Assessment**: AWS Client VPN now supports device posture assessment, integrating with CrowdStrike, Jamf, or JumpCloud to verify that connecting user devices meet your security and compliance requirements before granting network access. This is a significant security enhancement, implementing Zero Trust principles by ensuring devices meet security baselines (e.g., encryption status, risk level) using Cedar policies. It continuously re-evaluates posture, disconnecting non-compliant devices. This is critical for compliance with various regulatory frameworks requiring endpoint security. *Terraform Support*: Supported via `aws_ec2_client_vpn_endpoint` in [Terraform AWS Provider: aws_ec2_client_vpn_endpoint](https://registry.terraform.io/providers/hashicorp/aws/latest/docs/resources/client_vpn_endpoint) in **v5.68.0+** (latest **v5.72.0**).

## Announcements in this summary

- [AWS Control Tower](https://aws.amazon.com/about-aws/whats-new/2026/09/aws-control-tower-aft/) (Management and Governance)
- [Amazon Bedrock](https://aws.amazon.com/about-aws/whats-new/2026/10/amazon-bedrock-glm-5-3/) (Machine Learning)
- [AWS Batch](https://aws.amazon.com/about-aws/whats-new/2026/10/aws-batch-job-cloudwatch-metrics/) (Compute)
- [AWS Certificate Manager](https://aws.amazon.com/about-aws/whats-new/2026/10/AWS-Certificate-Manager-ACME-Privatelink) (Security, Identity, & Compliance)
- [Amazon EC2](https://aws.amazon.com/about-aws/whats-new/2026/10/ec2-ami-shared-tags) (Compute)
- [AWS Client VPN](https://aws.amazon.com/about-aws/whats-new/2026/10/aws-client-vpn-device-posture/) (Networking)
- [Amazon EKS](https://aws.amazon.com/about-aws/whats-new/2026/10/eks-auto-mode-advanced-compute-config/) (Containers)
- [Amazon Bedrock](https://aws.amazon.com/about-aws/whats-new/2026/10/amazon-nova-2.5-sonic/) (Machine Learning)
