Salesforce 利用 SageMaker 推理组件实现多可用区高可用
原标题：Spreading the load: How Salesforce met Multi-AZ HA with SageMaker Inference Components
AI 摘要
Salesforce 在构建 Agentforce 时，面临 Amazon SageMaker AI 推理组件（IC）默认放置策略无法满足多可用区（Multi-AZ）高可用合规要求的问题。AWS 通过新增 SchedulingConfig 参数（含 AvailabilityZoneBalance 和 PlacementStrategy）实现了 IC 副本跨 AZ 和实例的精细放置控制。Salesforce 采用 SPREAD 策略优先故障隔离，实现了 8 倍基础设施成本降低，同时满足 2-AZ 合规要求。
正文节选
Spreading the load: How Salesforce met Multi-AZ HA with SageMaker Inference Components When Salesforce set out to make Agentforce (Salesforce’s AI foundation for agents) highly available (HA) across multiple Availability Zones (AZs), the team faced a gap. Amazon SageMaker AI Inference Components (ICs) could cut GPU costs, but their default placement didn’t guarantee the Multi-AZ resilience Salesforce’s compliance bar required. For Salesforce, the ICs delivered an 8x reduction in infrastructure c