Stragglers, Not Failures: How Adaptive Hedged Requests Reduce p99 Latency by 74 Percent
InfoQ3782 字 (约 16 分钟)
92
Adaptive hedged requests reduce p99 latency by 74% by dynamically triggering hedges based on real-time latency distribution learning—not static thresholds or retries; DDSketch enables O(1) memory quantile estimation, paired with token-bucket rate limiting to prevent load amplification.
入选理由：在100个下游服务、各1%慢请求率的扇出架构中，63%的顶层请求会被至少一个慢请求拖累，导致单服务健康指标失真。
FeaturedArticle#Distributed Systems#Latency Optimization#Hedged Requests#DDSketch#Microservices英文