tail latency

Definitions

1. The response time experienced by the slowest requests — typically measured at the 99th or 99.9th percentile (p99/p999). In distributed systems, tail latency is amplified because a single slow component can delay an entire fan-out request. Google's research showed that as systems scale, even rare slow requests become frequent enough to affect most user-facing operations.

   In plain English: How slow your slowest requests are — even if most requests are fast, the unlucky 1% who get a slow response can ruin the overall user experience.

   Example: Our median latency is 20ms but the p99 is 3 seconds — that means 1 in 100 users waits 150x longer than average. The tail latency is killing our user experience.

Related Terms

• Latency
• Throughput
• p-value
• Tail Latency