Skip to content

Benchmark orchestration for Linux nodes

Linux Benchmark Library

Run repeatable workloads, collect multi-level metrics, and generate reports with a clean CLI and stable Python APIs.

Get started API reference
Quick run
lb config init -i
lb plugin list --enable stress_ng
lb run --remote --run-id demo-run

Provisioned runs are available in dev mode: --docker or --multipass.

Why this exists

Repeatable workloads

Standardize load patterns and run them across hosts or provisioned targets.

Layered architecture

Runner, controller, app, and UI are cleanly separated to keep coupling low.

Actionable artifacts

Raw metrics, journals, reports, and exports are organized per run and host.

Extensible plugins

Add new workloads via entry points and a user plugin directory.

Core layers

Layer Responsibility
lb_runner Execute workloads and collect metrics on a node.
lb_controller Orchestrate remote runs via Ansible and manage state.
lb_app Stable API for CLIs/UIs and integrations.
lb_ui CLI/TUI implementation.
lb_analytics Reporting and post-processing.
lb_provisioner Docker/Multipass helpers for the CLI.

Where to go next