aiengineering.guideaiengineering.guide

Glossary

tail latency

The slow end of the latency distribution (e.g. p99) — what the unluckiest requests see.

The high percentiles of the latency distribution — p95, p99, p99.9 — describing the requests that wait longest. Confusions: the mean hides the tail; a healthy average can coexist with a p99 that violates an SLO.

Also: p99 latency, p99

Enter to go · Esc to close