请问HN：在负载激增时，您是如何保持语音AI延迟低的？
大家好，
我正在进行一项研究，有一个问题想请教大家。当语音人工智能代理的流量激增时，你们通常会怎么处理？据我所知，Kubernetes 的原生自动扩缩容（HPA）往往无法及时跟上，导致延迟过高。如果你们愿意分享经验，我将非常感激。
谢谢！
查看原文
Hey guys,<p>I'm doing a research and have a question. What do you do when traffic to voice AI agents spikes? As far as I know, native autoscaler of Kubernetes (HPA) doesn't catch up quite often with that - resulting into a prohibitively high latency. Would be glad to know your experience if you don't mind.<p>Thanks!