Load testing
Restura has a built-in load runner for quick performance checks on a single HTTP request — no extra tooling, no separate config file. It reuses the exact same send path as a normal request, so auth, variable resolution, and the proxy round-trip all match what you’d get manually.
Running a load test
Section titled “Running a load test”- Open the HTTP request you want to hammer.
- Press ⌘K / Ctrl+K and choose “Run load test on current request”.
- Set two numbers and hit Run:
- Total requests — how many times to fire (default 100).
- Concurrency — how many in flight at once (default 10).
Results land in the Runs panel, where each run is recorded so you can compare and re-run.
What you get back
Section titled “What you get back”| Metric | Meaning |
|---|---|
| Req/s | Throughput across the whole run. |
| Mean / Min / Max | Latency summary in ms. |
| p50 / p90 / p95 / p99 | Latency percentiles, also drawn as a distribution bar chart. |
Each iteration is recorded as a sample with its status code and outcome, so failures show up in the success rate rather than skewing the latency numbers silently.
When to reach for the CLI instead
Section titled “When to reach for the CLI instead”This runner is for fast, interactive checks on one request. For repeatable, CI-gated runs across a whole collection — with JUnit / JSON / HTML reports — use @restura/cli.
Related
Section titled “Related”- HTTP / REST — load testing operates on HTTP requests.
- CLI runner — headless collection runs for CI.