As a Cloud Security Architect, DevSecOps Lead, and Terraform Specialist, I'm pleased to present today's critical AWS updates. These enhancements span compute, container orchestration, AI/ML, and data transfer, bringing significant improvements in performance, resilience, and security posture. Crucially, we'll examine their immediate and pending Terraform availability, ensuring our Infrastructure as Code (IaC) practices remain robust and up-to-date.
Compute Innovations & Security Enhancements
- Amazon EC2 C8g Instances Expand Regional Availability: AWS has expanded the availability of Amazon EC2 C8g instances, powered by Graviton4 processors, to AWS Asia Pacific (Taipei, New Zealand) and AWS GovCloud (US-East) regions. These instances deliver up to 30% better performance over Graviton3-based instances, offering enhanced compute, memory, and networking capabilities. From a security and architectural perspective, Graviton processors provide a strong foundation due to their custom silicon design, which can reduce the attack surface. Their availability in GovCloud regions is vital for US government workloads, ensuring compliance with stringent regulatory requirements. This expansion enables deploying high-performance, cost-effective, and secure compute closer to users in these strategic regions. *Terraform Support*: Supported via the
aws_instanceresource in Terraform AWS Provider: aws_instance in v5.0+. - Amazon ECS Introduces Early Success Criteria for Rolling Deployments: Amazon ECS now allows configuring 'Early Success Criteria' for rolling deployments, enabling users to define a 'healthy percent' of desired tasks that must be running and healthy before a deployment is marked successful. This significantly reduces deployment times and allows subsequent operations to proceed faster. Additionally, users can choose between 'BLOCKING' (cleanup completes before success) or 'DEFERRED' (asynchronous cleanup) for source service revision cleanup. This feature enhances deployment reliability and speed, reducing exposure windows during updates. 'BLOCKING' cleanup is a strong security practice, ensuring all old resources are fully terminated, preventing potential resource leakage or zombie tasks. This directly improves DevSecOps pipelines by enabling faster, more resilient deployments with configurable safety nets. *Terraform Support*: Supported in the
aws_ecs_serviceresource in Terraform AWS Provider: aws_ecs_service in v5.25.0. - Amazon EC2 P6-B200 Instances Now Available in Asia Pacific (Hyderabad): Amazon EC2 P6-B200 instances, accelerated by NVIDIA Blackwell GPUs, are now available in the AWS Asia Pacific (Hyderabad) Region. These instances offer up to 2x performance compared to P5en instances for AI training and inference, featuring 8 Blackwell GPUs with 1440 GB of high-bandwidth GPU memory and 5th Generation Intel Xeon processors. Built on the AWS Nitro System, they provide a secure and scalable foundation for AI workloads within Amazon EC2 UltraClusters. The Nitro System's hardware-level isolation, encrypted EBS volumes, and secure boot capabilities are critical for securing sensitive AI training data and models. This expansion brings cutting-edge GPU acceleration to a key APAC region, allowing customers to run large-scale, secure AI workloads closer to their data. *Terraform Support*: Supported via the
aws_instanceresource in Terraform AWS Provider: aws_instance in ~> 4.x and ~> 5.x. - Amazon EC2 C9g and C9gd Instances Powered by AWS Graviton5 Processors Now Available in Asia Pacific (Tokyo): AWS has launched C9g and C9gd instances, powered by AWS Graviton5 processors, in the Asia Pacific (Tokyo) region. These instances offer up to 25% better compute performance than Graviton4-based instances and are the first to feature the Nitro Isolation Engine, which uses formal verification to provide mathematical assurance of workload isolation. C9gd instances also include local NVMe-based SSD storage. The Nitro Isolation Engine represents a significant security advancement, enhancing the confidentiality and integrity of customer data in multi-tenant environments. This expansion provides secure, high-performance, and cost-effective compute options for HPC, batch processing, and ML inference in a critical APAC region. *Terraform Support*: Supported via the
aws_instanceresource in Terraform AWS Provider: aws_instance in v5.50.0 (first) to v5.70.0+ (latest). - Amazon ECS Managed Daemons Can Now Be Designated as Non-Critical: Amazon ECS now allows designating Managed Daemons as 'non-critical'. If a non-critical daemon task fails, stops, or becomes unhealthy, the underlying container instance remains active, existing application tasks continue to run, and new application tasks can still be placed. This ensures immediate task launches even if a non-critical daemon fails to start. ECS also emits EventBridge events and records service action logs for enhanced observability. This feature significantly improves the resilience and availability of ECS services by preventing non-critical daemon failures from disrupting core application functionality. From a DevSecOps perspective, this allows for more robust application design and graceful degradation, while enhanced logging aids in monitoring security posture and detecting anomalies. *Terraform Support*: Pending provider release (use AWS CLI/SDK or custom provisioner in the interim).
- Amazon EC2 M9g and M9gd Instances Powered by AWS Graviton5 Processors Now Available in Four Regions: AWS has made M9g and M9gd instances, powered by Graviton5 processors, available in Europe (Ireland) and Asia Pacific (Singapore, Sydney, Tokyo) regions. These instances offer up to 25% better performance than Graviton4-based instances for memory-intensive workloads and are also built on the 6th-generation AWS Nitro System, featuring the Nitro Isolation Engine with formal verification. M9gd instances include local NVMe-based SSD storage. Similar to the C9g/C9gd, the Graviton5 processors and Nitro Isolation Engine provide a robust foundation for secure and performant memory-intensive workloads, critical for databases, in-memory caches, and big data analytics. Expanding availability to these regions offers more options for deploying secure, high-performance, and cost-effective solutions closer to users and data, aiding regional compliance. *Terraform Support*: Supported via the
aws_instanceresource in Terraform AWS Provider: aws_instance in v5.60.0 (first) to v5.70.0+ (latest). - Amazon EC2 Enables AMI Owners to Define Instance Type Compatibility: AMI owners can now define which instance types are compatible or incompatible with their AMIs, with any launch attempt on a non-permitted instance type automatically blocked. By default, AMIs can be launched on any instance type until restrictions are explicitly applied. This is a critical security and operational control, allowing AMI owners to enforce best practices by restricting AMIs to instance types that meet specific performance, security, or compliance requirements (e.g., FIPS-validated instances). This reduces misconfiguration risks, ensures consistent deployments, and helps maintain compliance by preventing unauthorized instance type usage. *Terraform Support*: Supported via the
aws_amiresource in Terraform AWS Provider: aws_ami in v5.60.0.
Developer Tools & Observability
- AWS MCP Server Adds Serverless Capability for Lambda Function Diagnosis: The AWS MCP Server now offers a serverless capability, enabling coding agents like Claude Code and Kiro to efficiently diagnose issues with Lambda functions and their connected resources across various AWS services (API Gateway, EventBridge, S3, DynamoDB, SNS, SQS, Step Functions). The agent correlates error signals against a 7-day baseline, identifies trends, retrieves deployed configurations, provides change timelines, and analyzes service latency, consuming fewer tokens by consolidating API calls. This powerful diagnostic tool centralizes data, helping detect anomalous behavior or misconfigurations that could indicate security incidents. The ability to retrieve configurations and change timelines is invaluable for incident response and auditing. However, strict IAM controls and least privilege principles are essential when granting agents broad diagnostic access. *Terraform Support*: Pending provider release (use AWS CLI/SDK or custom provisioner in the interim).
Migration & Transfer Security
- AWS Transfer Family SFTP Connectors Support Credential Rotation: AWS Transfer Family SFTP Connectors now continue running file transfers seamlessly while you rotate credentials for remote SFTP servers. Connectors can retrieve credentials from an ordered list of AWS Secrets Manager version stages (e.g., current, previous), automatically trying each version until one succeeds. This is a significant security enhancement, as automated and seamless credential rotation is a best practice for reducing the risk of compromised credentials. Integration with AWS Secrets Manager centralizes secret management and allows for robust rotation policies, improving the resilience and security posture of data transfer workflows by eliminating downtime and reducing the window of vulnerability for static credentials. *Terraform Support*: Supported in the
aws_transfer_connectorresource in Terraform AWS Provider: aws_transfer_connector in v5.65.0 (first) to v5.72.0 (latest).
Machine Learning Performance
- Amazon SageMaker AI Batch Transform Now Supports G6e Instances: Amazon SageMaker AI Batch Transform jobs now support Amazon EC2 G6e instances, featuring NVIDIA L40S Tensor Core GPUs with 48 GB of memory and third-generation AMD EPYC processors. This enables more efficient execution of offline inference for demanding AI models, including large language models (LLMs) and generative diffusion models. From a security standpoint, leveraging these powerful instances within a secure SageMaker environment (VPC, IAM roles, data encryption at rest/in transit) is paramount. The ability to process large models offline can reduce real-time latency concerns and allow for more thorough security scanning of model outputs before deployment. *Terraform Support*: Supported via the
instance_typeargument of theaws_sagemaker_transform_jobresource in Terraform AWS Provider: aws_sagemaker_transform_job in v2.1.0 (resource introduction) and all subsequent v5.x.x versions.