| p50 Latency | Median response time — 50% of requests completed faster than this. A good baseline for typical user experience. |
| p95 Latency | 95% of requests completed faster than this. Captures the slower tail; useful for SLA definitions. |
| p99 Latency | 99% of requests completed faster than this. Represents near-worst-case response times experienced by users. |
| p50 Completion | Elapsed time from test start by which 50% of all requests had finished. Accounts for both when the request was sent and how long it took. |
| p95 Completion | Elapsed time from test start by which 95% of all requests had finished. |
| p99 Completion | Elapsed time from test start by which 99% of all requests had finished. High values indicate request queuing or late-stage slowdowns. |
| p50 done / p95 done / p99 done | Per-endpoint completion time percentiles — same as above but scoped to a single endpoint's requests. |
| OK% | Percentage of requests that returned a 2xx HTTP status with no network error. |
| Avg RPS | Total requests divided by total test duration — the overall throughput of the run. |
| Errs | Count of failed requests (non-2xx status or network errors such as timeout, connection refused). |