{
  "id": 20782420,
  "title": "Keep inference running through regional capacity shortfalls with cross-Region inference routing in Amazon SageMaker AI",
  "url": "https://urgent.news/2026/11/19/keep-inference-running-through-regional-capacity-shortfalls-with",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-19T18:20:54.000Z",
  "source": {
    "name": "AWS Machine Learning",
    "slug": "aws-machine-learning",
    "url": "https://aws.amazon.com/blogs/machine-learning/keep-inference-running-through-regional-capacity-shortfalls-with-cross-region-inference-routing-in-amazon-sagemaker-ai/"
  },
  "original_language": "en",
  "account": "Amazon SageMaker AI now offers cross-Region inference routing, simplifying serving models across multiple AWS Regions. Using a new resource called the inference router, developers can route requests to the most appropriate endpoint without having to build their own cross-Region routing system. The router keeps requests in the local Region when healthy, shifts traffic to another destination when needed, and gradually returns traffic after recovery.\n\nThe system evaluates destination health based on recent requests, with server errors (5xx), throttles, and timeouts lowering the health score. Model latency isn't considered a factor, so successful responses keep the endpoint serving traffic. The router operates in the same Region as the endpoints and passes relevant information, such as InferenceId and CustomAttributes, through to the endpoints. Each destination must already be running in its Region, and the router doesn't create capacity, deploy models, or copy model artifacts.\n\nTo implement cross-Region inference routing, developers choose which endpoints and inference components to use for each model, then register destinations under routing targets. The router prefers destinations in the local Region and spreads requests evenly among healthy local destinations. When local destinations can't serve reliably, traffic moves to other destinations in the same Region. The router manages failover thresholds and timing, so no tuning is required. If health information for remote destinations becomes stale, the router keeps traffic in the local Region to avoid gaps in monitoring.",
  "summary": "Cross-Region inference routing in Amazon SageMaker AI gives your application one invocation path across the endpoints and inference components you already run, in one or more AWS Regions and AWS accounts. The router prefers its own Region and shifts new requests to another destination only when the local one can't serve reliably.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}