Understanding Server Load
How we calculate. Load % = (used ÷ max) × 100 for one capacity dimension. The form uses the same arithmetic as the worked examples on this page. See our methodology and accuracy policy.
Real-world scenario: A typical Server Load case uses used capacity 7.2 and max capacity 8. Enter the same figures below to reproduce the worked path.
What is Server Load?
A utilization-style metric for infrastructure under test. State which resource you measured—CPU load ≠ memory load.
- One resource per calculation
- >85–90% sustained often means little headroom
- Compare to latency as load rises
The Formula
Worked Example
Common Use Cases
- Capacity headroom: see remaining room under peak
- Autoscale triggers: map % to policies
- Perf bottlenecks: correlate load with latency
Pro Tips
- Use steady-state averages, not single spikes alone
- Track p95 load for bursty systems
- Document the unit in every report
Limitations: Server Load figures are educational performance-testing aids—not contractual capacity guarantees. Validate with calibrated load tools and production telemetry.
FAQ
Is this CPU load average?
It can be—if your used/max inputs are CPU capacity. The formula is generic for any capacity pool.
Load > 100%?
Means used exceeded the defined max (overcommit or bad max). Investigate before trusting the figure.
Authoritative References
For performance and quality engineering context, consult:
- ISTQB — performance testing syllabus topics
- W3C Navigation Timing — browser performance timing
- CMG / capacity forums — capacity and performance practice communities