dnsperf — Measure Resolver Throughput and Latency
dnsperf drives a controlled query load at a name server and measures what comes back: sustained queries per second, response latency, and the point at which the server starts dropping or delaying answers. It replays a query file at a rate you choose, so the load resembles real traffic patterns rather than a single query repeated.
Why run it
DNS capacity is invisible until it is not. A resolver that answers instantly under normal load can fall off sharply past a threshold nobody has measured, and when it does, the symptom is not a DNS error — it is every dependent service becoming slow simultaneously, which sends people looking anywhere but at DNS. Measuring the ceiling deliberately, before a traffic spike finds it, is the difference between a known capacity limit and a mystifying incident.
Why automate it
A benchmark is a point in time, and the number that matters is the trend. Resolver capacity degrades as zones grow, as recursion depth increases, and as the host picks up other work; a figure measured at commissioning may be badly wrong a year later. Running the benchmark on a schedule turns a one-off measurement into a capacity trend, which is what lets you plan an upgrade rather than react to a saturation event.
What it gives you
A real number for a limit most teams only ever estimate. Knowing where your DNS capacity ends — and watching that number over time — turns a class of confusing, whole-platform slowdowns into something you can anticipate.