Performance Studio

Abstract

Performance Studio is a built-in workbench for authoring and running load/performance tests powered by k6. It lets you turn a recorded browser session or an existing API test case into a k6 script, run it with a configurable load profile, and review the results — all without leaving the IDE.

Prerequisite: k6

Performance Studio drives the open-source k6 load-testing engine. Install k6 and make sure it is available on your PATH before using ingenious perf run/validate. If k6 cannot be found, both the CLI and the IDE panel will show a warning with an install hint.

Download k6 from k6.io/docs/get-started/installation.

Opening Performance Studio

Click the Performance Studio button (speedometer icon) on the main toolbar.

Performance Studio opens as its own panel, next to API Workbench:

Performance Studio panel

Project Folder Layout

Performance Studio artifacts live alongside the rest of your project, under dedicated Performance and Results folders:

<Project>/
├── Performance/
│   ├── scripts/        # generated/hand-edited k6 scripts (.js)
│   ├── profiles/        # load profiles (YAML) — VUs, stages, thresholds
│   ├── recordings/       # recorded browser traffic (.har)
│   └── rules/          # auto-correlation rules extracted from recordings
└── Results/
    └── Performance/
        └── <script>/
            └── <timestamp>/   # run output: summary, logs, k6 dashboard report

Workflow

The typical Performance Studio workflow follows these stages:

Record (optional) → Export → Validate → Run → Report

1. Record browser traffic (optional)

If you don't already have an API test case to export from, record a browsing session to a HAR file:

WindowsMac/Linux

ingenious perf record https://example.com --project "Projects\Demo"

./ingenious.command perf record https://example.com --project "Projects/Demo"

Option	Description
`--out`	HAR file path (default: `Performance/recordings/<host>_<timestamp>.har`)
`--duration`	Stop automatically after N seconds (default: interactive — press Enter to stop)
`--headless`	Record without a visible browser window

The recorder opens a real browser, captures every request/response, and saves it as a .har file under Performance/recordings/.

2. Export a k6 script

Generate a k6 script either from an existing API test case or from a HAR recording:

WindowsMac/Linux

ingenious perf export Demo\Login\LoginSmoke --type http --profile smoke

./ingenious.command perf export Demo/Login/LoginSmoke --type http --profile smoke

Option	Description
`--type`	Script flavor: `http` (protocol-level, default) or `browser`
`--profile`	Load profile baked into the generated script's `options` (default: `smoke`)
`--out`	Output script path (default: `Performance/scripts/<name>.js`)
`--force`	Overwrite a hand-edited script
`--url-filter`	HAR exports only — keep requests whose URL contains this text
`--include-static`	HAR exports only — keep static assets (CSS/JS/images)
`--auto-correlate`	HAR exports only — detect tokens from a response that reappear in a later request, save the correlation rules to `Performance/rules/`, and apply them automatically
`--rules`	Rules file to apply (default: `Performance/rules/<script>.rules.yaml`)

Two export sources

<Project>/<Scenario>/<TestCase> — flattens the test case (inlining any Execute/User Intent steps) and turns its API steps into a k6 protocol script.
A .har file path — turns a recorded browsing session into a k6 script (--type browser for a Playwright-driven k6 browser script, --type http for a protocol-level replay of the captured requests).

3. Validate

Before running a full load test, do a quick 1 VU / 1 iteration debug run with a full HTTP trace:

WindowsMac/Linux

ingenious perf validate LoginSmoke.js --full

./ingenious.command perf validate LoginSmoke.js --full

Use --full to print the entire trace instead of just the tail — useful for diagnosing a failing correlation rule or an unexpected response.

4. Run

WindowsMac/Linux

ingenious perf run LoginSmoke.js --vus 20 --duration 2m

./ingenious.command perf run LoginSmoke.js --vus 20 --duration 2m

Option	Description
`--vus`	Override the number of virtual users baked into the script's load profile
`--duration`	Override the duration (e.g. `30s`, `2m`)
`--detach`	Start the run in the background and return immediately; control it afterwards with `perf status`/`logs`/`scale`/`cancel`
`--dashboard`	Enable the k6 web dashboard (live graphs in the browser, plus an HTML report export); implies `--detach`

In the IDE, the panel shows the run live: a status pill (idle/running/completed/failed), a live VUs / requests-per-second chart, and a streaming output console — all without needing the CLI.

Performance Studio live run

Managing a detached (background) run

Command	Description
`ingenious perf status [runId]`	Live status + metrics of an async run (defaults to the newest running one)
`ingenious perf logs [runId] [--lines <n>]`	Output of an async run (tails the last `n` lines, default `40`)
`ingenious perf scale <vus> [runId]`	Change the VU count of a running k6 test
`ingenious perf cancel [runId]`	Stop an async run (graceful, then kill)

5. Report

WindowsMac/Linux

ingenious perf report latest

./ingenious.command perf report latest

Variant	Description
`report latest` (default)	Show the latest run's summary
`report history`	List previous runs for the current project
`report compare <baselineRunId> <candidateRunId>`	Compare metrics between two runs
`report junit [runId] --out <file>`	Export a run as a JUnit XML report (useful in CI), default stdout

Performance Studio report view

Load Profiles

A load profile is a reusable YAML definition of VUs/stages/thresholds that gets baked into an exported script via --profile.

Command	Description
`ingenious perf profile list`	List built-in and project load profiles
`ingenious perf profile show <name>`	Show a profile as YAML
`ingenious perf profile create <name> [--from <template>] [--force]`	Create a new project profile by copying a built-in template (`--from`, default `smoke`). Project profiles are stored under `Performance/profiles/`

Listing Artifacts

ingenious perf list [scripts|profiles|recordings|runs]

Lists the scripts, profiles, recordings, or past runs available for the current project.

Quick Reference

Command	Purpose
`ingenious perf record <url>`	Record browser traffic to a HAR file
`ingenious perf export <target>`	Generate a k6 script from a test case or a `.har` file
`ingenious perf validate <script>`	Debug run: 1 VU, 1 iteration, full HTTP trace
`ingenious perf run <script>`	Execute a k6 script (load run)
`ingenious perf status`/`logs`/`scale`/`cancel`	Manage a detached (`--detach`) run
`ingenious perf report [latest\\|history\\|compare\\|junit]`	Show performance run results
`ingenious perf profile list\\|show\\|create`	Manage load profiles
`ingenious perf list [scripts\\|profiles\\|recordings\\|runs]`	List Performance Studio artifacts

Note

ingenious perf can also be invoked with its alias ingenious performance. Append --help to any subcommand to see its full option list.