Load testing

Local, process-bound virtual users (VUs) against a Signoff or Postman collection. Available in the Runner Load panel and via

npm run runner -- load

--max-vus

Load panel (UI)

• Platform Availability: Full Parity
• Overview: Configure VUs, steady-state duration, ramp-up/down, think time, and pass/fail thresholds (p95, error rate) against the current Runner collection/folder scope. Results land as a load JSON report plus on-screen metrics.
• How It Works (Windows App):
  1. Open Runner, select collection/folder and environment.
  2. Open the Load panel; set VUs, duration, ramp, think time, thresholds.
  3. Start the load session; watch live metrics; review the JSON report when finished.

• How It Works (VS Code Extension):
  1. Same Load panel inside the shared Runner shell.
  2. Keep the webview open for the full duration — closing the panel stops the session host path.
  3. Prefer CLI load in CI for headless runs (same engine).

• Code / Configuration Examples:

npm run runner -- load \
  -c ./Pets.signoff.json \
  -e ./staging.signoff.env.json \
  --vus 20 \
  --duration 60000 \
  --ramp-up 10000 \
  --ramp-down 5000 \
  --think-time 200 \
  --p95 500 \
  --error-rate 0.01 \
  --max-vus 50 \
  -o ./reports \
  -v

• Edge Cases, Troubleshooting & Pro-Tips:
  • Exit codes: 0 thresholds pass, 1 fail, 2 usage/load/validation error.
  • High VUs on a laptop will bottleneck the client CPU before the server — raise gradually.

  • --users

    /

    --duration-ms

    are aliases for

    --vus

    /

    --duration

    .

CLI load subcommand

• Platform Availability: Full Parity (CLI on any Node host with the backend package)
• Overview: Headless load sessions for local perf checks and CI gates. Reports write to

  --output

  (default cwd). Variable overrides and folder/request filters match the

  run

  command style.
• How It Works (Windows App):
  1. Install/use the backend package that ships

     npm run runner

     .
  2. Export or point at collection + environment files on disk.
  3. Invoke

     load

     with thresholds; fail the pipeline on exit code 1.

• How It Works (VS Code Extension):
  1. Use the same CLI from a terminal (extension UI is optional).
  2. Resolve workspace-relative paths to absolute files before scripting.
  3. Do not rely on the webview being open for CLI load.

• Code / Configuration Examples:

-c, --collection <path>

-e, --environment <path>

--env-file <path>

-o, --output <dir>

cwd

--vus <n>

10

--users <n>

--vus

--duration <ms>

30000

--duration-ms <ms>

--duration

--ramp-up <ms>

0

--ramp-down <ms>

0

--think-time <ms>

0

--p95 <ms>

--error-rate <rate>

--max-vus <n>

50

-v, --verbose

false

--var <pair...>

key=value

--folder <nameOrId>

--request <nameOrId...>

npm run runner -- load -c ./col.json --vus 10 --duration 30000 --p95 800 --error-rate 0.05

• Edge Cases, Troubleshooting & Pro-Tips:
  • Soft cap: requested VUs are clamped to

    --max-vus

    (default 50).
  • Combine

    --folder

    or

    --request

    to load a hot path without the whole collection.
  • Thresholds are evaluated on the aggregate session — not per individual request name unless you filter scope.

Thresholds & reporting

• Platform Availability: Full Parity
• Overview: Optional

  --p95

  and

  --error-rate

  turn a load session into a gate. The primary artifact is a load JSON report under

  --output

  .
• How It Works (Windows App): Set the same fields in the Load panel; open the report folder when the run completes.
• How It Works (VS Code Extension): Same panel fields; or run CLI and open the JSON in the editor.

• Code / Configuration Examples:

# Fail CI if p95 > 500ms or more than 1% errors
npm run runner -- load -c ./smoke.signoff.json --vus 15 --duration 45000 --p95 500 --error-rate 0.01
echo Exit: $?

• Edge Cases, Troubleshooting & Pro-Tips:
  • Cold starts skew p95 — use

    --ramp-up

    or a short discard window operationally.
  • Error rate is 0–1 (

    0.01

    = 1%), not a percentage integer.

Related

• Runner
• Monitors & scheduler
• MCP & CLI