RTE logo
← Back to Node.js performance map
Node.js Performance

Percentile Latency

Measure tail latency. Average hides the slow users.

p50/p95/p99

Detailed Description

Percentile latency tells you what real users experience at different points in the latency distribution. P50 shows the typical request, while P95 and P99 expose the slow tail that averages hide.

In Node.js APIs, tail latency usually comes from event-loop blocking, slow database calls, cold caches, overloaded downstream services, or too much concurrency. Track percentiles per endpoint, not only globally.

Mental Model

request start
-> response finish
-> store duration
-> p50 / p95 / p99

When To Use

  • ->Define API SLOs using p95 or p99 latency
  • ->Find tail-latency regressions after deploys
  • ->Compare before/after optimization work
Tools: perf_hooks, process.hrtime.bigint, prom-client histogram
percentiles.js

Live Editor

Loading...

Output

Run code to see output...

Comments

Sign in to leave a comment. Your name and photo come from Google; nothing else is shared.

Loading comments...