resperf — Find the Point Where a Resolver Saturates
Where dnsperf holds a steady rate, resperf ramps the load continuously upward and records the point at which the server stops keeping up. The output is a curve rather than a single figure: response rate and latency plotted against offered load, with a clearly visible knee where the server transitions from comfortable to overwhelmed.
Why run it
Because the knee is the number you actually need. Knowing a resolver handles a given rate comfortably does not tell you how much headroom exists, and headroom is what determines whether a traffic spike is absorbed or becomes an outage. The shape of the curve is informative too: a server that degrades gracefully is a different operational proposition from one that falls off a cliff, and the plan for each is different.
Why automate it
Headroom shrinks quietly. Every new zone, every additional record, every extra service sharing the host consumes some of it, and none of those changes announces its effect on capacity. Re-running the ramp on a schedule tracks the knee over time, which is the only way to notice that the margin you sized for eighteen months ago has been eaten. It also validates that a capacity change you made actually moved the number.
What it gives you
Your true headroom, expressed as a number rather than a hope, and tracked as it changes. That is what makes DNS capacity a planning decision instead of a post-incident discovery.