Automation / DevOps / IaC

Amazon SageMaker Unified Studio now allows provisioning via Terraform using the open-source terraform-aws-sagemaker-unified-studio module. This enables platform teams to integrate SageMaker domains into existing infrastructure-as-code pipelines with version-controlled templates. The update supports consistent deployment across development, staging, and production accounts for shared AI and data workflows.

• Use the open-source terraform-aws-sagemaker-unified-studio module for provisioning
• Integrate SageMaker Unified Studio domains into existing IaC pipelines
• Maintain environment consistency across dev, staging, and prod accounts
• Provision managed workspaces with version-controlled templates
• No direct database engine changes; focus is on platform provisioning

AWS Certificate Manager now exposes a fully managed ACME server endpoint, allowing standard clients like Certbot and cert-manager to issue 45-day public TLS certificates from Amazon Trust Services. This feature addresses the upcoming CA/Browser Forum mandate for shorter certificate lifetimes by enabling standardized, automated issuance and renewal workflows. Administrators can enforce governance through domain scoping and wildcard policies while delegating requests to application teams.

• Provision public TLS certs using any ACMEv2 client like Certbot or cert-manager directly from ACM.
• Certificates issued via ACM have a 45-day validity to align with industry security trends.
• Governance controls allow restricting domain scopes and wildcard usage per client.
• Automates renewal processes, reducing manual overhead as 47-day limits approach in 2029.
• Enables delegation of certificate requests to app teams without distributing private CA keys.

Firecrawl is an open-source API designed for large-scale web search and scraping, specifically optimized for AI agents. It extracts content from JS-heavy pages and converts it into clean Markdown or structured JSON, handling infrastructure challenges like rotating proxies and rate limiting. The service claims industry-leading reliability across 96% of the web with a P95 latency of 3.4 seconds.

• Provides LLM-ready output formats including clean Markdown and structured JSON.
• Handles complex scraping infrastructure like rotating proxies and orchestration automatically.
• Optimized for real-time agents with P95 latency of 3.4s across millions of pages.
• Covers 96% of the web, including JavaScript-heavy dynamic content pages.
• Available as both open-source software and a hosted service.

AWS has introduced Service Events for CloudWatch Application Signals, enabling automatic capture of exception and latency snapshots, function-level metrics, and deployment events. This feature requires no additional code changes, relying instead on existing instrumentation via ADOT SDKs or the CloudWatch Observability EKS add-on. Engineers can now rapidly diagnose whether recent deployments introduced new exceptions by navigating directly to the Errors section in the console.

• Service Events automatically captures error and latency snapshots without code changes.
• Deployment events are tracked alongside performance data for faster root cause analysis.
• Requires existing instrumentation via ADOT SDKs or CloudWatch Observability EKS add-on.
• Access captured data directly in the CloudWatch console under Service > Errors.

Amazon SageMaker HyperPod now tracks AMI versions across clusters to detect drift and automatically applies security patches without interrupting long-running training jobs. This update removes the manual burden of patching and reduces the risk of breaking bundled software like NVIDIA drivers or CUDA during updates. The feature provides cluster administrators with clear visibility into the specific AMI versions currently in use.

• Eliminates manual security patching for long multi-day training workloads.
• Provides visibility into AMI versions to detect cluster drift quickly.
• Prevents accidental breaks to bundled software like CUDA drivers during updates.
• Reduces operational overhead for cluster administrators managing foundation models.