latency

The time delay between a request being sent and a response being received. In web applications, latency affects user experience: load times above 200ms become perceptible; above 1s they affect bounce rate and conversion. P99 latency (the worst 1% of requests) is often the operationally important metric — it captures the outlier experience that affects enterprise SLAs and the users most likely to complain.

When you'd see it: Performance monitoring dashboards, SLA negotiations, infrastructure optimization discussions.

Why it matters: Latency directly affects user experience, conversion rates, and SLA compliance. Google's core web vitals and Amazon's documented correlation between page load time and revenue both make the business case concrete.

Common mistakes: Optimizing average latency while ignoring tail latency. Mean latency can look acceptable while P99 latency is bad enough to cause user-visible problems for a significant fraction of sessions.

Related observability, CDN, SLO, rate limiting