Getting Started

Using the platform

A tour of the perfscale admin console — machines, configurations, tests, and results

The admin console is where you drive perfscale from a browser: connect load generators, define load shapes, author and run tests, and read results. Sign in at /admin on your instance (e.g. perfscale.su/admin). Want to look around first? The sign-in page has a Try the demo button that prefills a read-only demo workspace.

1. Machines — your load generators

A machine is any host running the perfscale agent: a bare-metal box, a VM, a container, or a cloud instance. The Machines page lists them with region, capacity (how many virtual users each can drive), status, and live latency.

The Machines page listing connected load generators

Add capacity with Add device — the panel at the bottom has one-line install snippets for Linux, macOS, and Windows, and for AWS, Azure, GCP, Docker, and Kubernetes. The agent registers itself and appears here within seconds. See Machines for the lifecycle and statuses.

2. Configurations — load shapes and variables

Configurations are reusable load profiles — how many virtual users, for how long, with what ramp. The built-in presets (debug, smoke, load, stress, spike, soak) match the CLI presets, and you can add custom profiles with ramp-up / steady-state / ramp-down phases.

The Configurations page with built-in presets and environment variables

The same page holds Environment Variables injected into every run — API_BASE_URL, secrets, connection strings. Mark a value masked to keep it out of logs and the UI.

3. Tests — author and run

A test pairs a script (k6, JMeter, Locust, or a native perfscale YAML step definition) with a configuration. The editor has Code, YAML, and Visual views; the right panel shows the environment variables in scope.

The Tests page with the script editor and environment variables

Hit Run (or Start Test to pick a machine and configuration) and the run is dispatched to an available machine. You can also Sync from Git to keep tests in version control. See Tests & Tasks for the test and task lifecycle.

4. Runs & logs — watch it execute

Every run streams its output to Runs & logs: dispatch, per-request lines, check results, and the final metric summary, all timestamped and filterable.

The Runs & logs page streaming execution output

Aggregated latency percentiles and throughput land on the Metrics dashboard, and the whole flow can be automated from Triggers (schedules and webhooks) or the REST API.

Next steps

Quick Start — run your first test.
Machines and Tests & Tasks — the core concepts.
FIX Protocol — load-test trading venues (Scale & Enterprise).