Resource Utilization Calculator

Utilisation: ρ = λ / μ

Resource utilisation is the fraction of time a server is busy: ρ = λ / μ, where λ is the arrival rate and μ is the service rate. Equivalently, ρ = busy time ÷ available time. ρ < 1 does not mean no queue — variability alone creates waiting time even at ρ = 0.7. Kingman's formula shows mean wait grows as ρ / (1 − ρ) — a hyperbolic explosion near ρ = 1. Engineering rule of thumb: target ρ ≤ 80% for production capacity, ≤ 70% for aggressive SLAs. Example: server busy 6 h of 8 h → ρ = 0.75 = 75%.

Applications

Web-server capacity planning (DevOps/SRE), database connection-pool sizing, Kubernetes HPA CPU utilisation thresholds, OS process scheduling, factory-floor OEE in Six Sigma, machine-shop loading, and theme-park ride throughput. Anything with a queue and a server uses ρ as its primary KPI.

FAQ

Why not run at 95%? Kingman: wait time roughly doubles from ρ = 0.8 to ρ = 0.9, then doubles again from 0.9 to 0.95. Tail latency explodes long before throughput does.

Multiple servers (M/M/c)? Per-server utilisation is ρ = λ / (c · μ). Queues still build when ρ > ~80% even with many servers.

What does ρ > 1 mean? Demand exceeds capacity — the queue grows unbounded. The system is unstable; you need more capacity or less arrival rate.