Load Balancer Target Capacity Calculator
Estimate how many application targets can serve a peak request rate while staying below a tested utilization ceiling.
Use per-target benchmark results and set a separate minimum for redundancy. Maintained by Buildopsy · Updated .
Target fleet scenario
Formula and assumptions
Required targets = ceil(peak RPS ÷ (tested RPS per target × utilization ceiling)). Final count is the larger of this value and the configured redundancy minimum.
This is a steady-state capacity estimate. It does not model target warm-up, health-check removal, uneven load distribution, autoscaling delay, connection reuse, payload mix, latency objectives, or provider-specific limits. Test with realistic traffic and ensure remaining targets can carry load during a failure.
Frequently asked questions
How many load balancer targets do I need?
Use measured sustainable RPS per target at your intended utilization limit, divide the peak traffic by that capacity, round up, then enforce your minimum redundancy count.
What should per-target capacity be based on?
A representative benchmark of the actual request mix, instance type, app configuration, and latency target. Provider limits and failover needs should be checked separately.
Related calculators
All engineering calculators · Kubernetes HPA replicas · Connection pool sizing · Queue backlog and drain time