AWS Weekly Roundup (September 28, 2026)

Amazon CloudWatch Omni

Now you can observe applications and AI agents together in a single collaborative experience. Built on OpenTelemetry, existing telemetry shows up with nothing to reconfigure, reachable through one URL with enterprise SSO — no console access needed. It auto-discovers services, maps dependencies, and brings the AWS DevOps Agent into investigation sessions to correlate signals and trace root causes.

Enhanced custom event buses in EventBridge

A new custom event bus purpose-built for organizations scaling event-driven applications across teams and accounts. Deploy a single centralized bus shared across every account through AWS RAM, with optional event ordering, a simplified Subscriber resource bundling filtering/targets/retries, content-based deduplication, and synchronous invocation for targets like Lambda. A new ingress/egress pricing model replaces compounding cross-account routing charges; existing buses keep working as "classic."

SageMaker HyperPod Inference Gateway

Front LLM inference on SageMaker HyperPod with a Kubernetes-native, GPU-aware routing layer that deploys as a single EKS managed add-on with zero application changes. Instead of round-robin, it routes on real-time inference signals — KV cache utilization, queue depth, prefix cache hits, predicted latency — cutting first-token latency by up to 82% in mixed-hardware and bursty scenarios. Works with any OpenAI-compatible model server including vLLM and SGLang.

AI agent skills for SES and End User Messaging

Build and send messages by asking your AI coding agent in plain language. Amazon SES and AWS End User Messaging publish AI agent skills for the AWS MCP Server, giving agents step-by-step validated guidance for tasks like verifying a sending identity, sending a production email, or building a branded RCS agent.

Bottom line

The HyperPod Inference Gateway is the standout: KV-cache-aware routing with zero app changes turns raw GPU nodes into latency-optimized inference infrastructure — a pattern every self-managed LLM team should steal the idea of.