Benchmarking
The benchmark tool runs a single request repeatedly and reports latency statistics. Use it to measure baseline performance, detect regressions, or stress-test an endpoint with concurrent requests.
Opening the Benchmark Tool
Section titled “Opening the Benchmark Tool”Right-click a saved request in the sidebar and select Benchmark, or open it from the Command Palette (Benchmark Request).
Configuration
Section titled “Configuration”| Option | Description | Default |
|---|---|---|
| Iterations | Total number of requests to send | 100 |
| Concurrency | Number of requests sent in parallel | 1 |
| Delay | Milliseconds to wait between iterations | 0 |
| Timeout | Per-request timeout in milliseconds | 30000 |
Concurrency 1 (sequential) measures baseline latency without contention. Higher concurrency simulates multiple simultaneous clients and reveals how the endpoint behaves under load.
Running
Section titled “Running”Click Run to start. A progress bar shows completed iterations. Click Cancel to stop early. Partial results are displayed for cancelled runs.
Results
Section titled “Results”After the run, the results panel shows:
| Metric | Description |
|---|---|
| Passed / Failed | Counts of successful and failed requests |
| Total time | Wall-clock time for the full run |
| RPS | Requests per second (throughput) |
| Min | Fastest iteration |
| Max | Slowest iteration |
| Mean | Average duration |
| Median (p50) | Middle value |
| p90 / p95 / p99 | Tail latency percentiles |
A distribution chart shows the spread of response times across all iterations.
Exporting Results
Section titled “Exporting Results”Click Export to save results as JSON. The export includes all per-iteration data: status, duration, and error (if any), plus the full statistics summary.
The benchmark tool is also available from the command line for CI integration. See CLI: Benchmark.