Skip to content

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.2.0]

Breaking

  • JSON output: sampling field removed from the output schema. Latency percentiles are now exact (computed from a full HDR histogram) — no approximation, no sampling state to report.
  • Config: result_buffer and sample_threshold config fields removed. The old reservoir sampling pipeline has been replaced; these settings have no effect and will cause a parse error if present.

Changed

  • Histogram-based statistics pipeline — exact latency percentiles for any run size, replacing the reservoir sampling approach
  • Worker-pool VU model with Arc-based clone reduction — lower memory overhead at high concurrency
  • CompiledTemplate struct — template compilation errors are now reported at startup, before any requests fire
  • Write-to-buffer renderer with pre-resolved globals — faster per-request body generation
  • Replaced expect/unwrap with a typed RunError — no more panics during a run, errors surface cleanly

Fixed

  • CLI reference: default value for --concurrency corrected to 10 (was incorrectly documented as 100)

[0.1.7]

Changed

  • Restore publish pipeline
  • Drop Windows from release targets
  • Bump CI dependencies: actions/setup-python, docker/login-action, docker/build-push-action

[0.1.5]

Changed

  • Improve documentation structure and content
  • Set custom domain lmn.talek.cloud for docs site
  • Fix cargo-deb release pipeline

[0.1.4]

Changed

  • Update docs URLs to lmn.talek.cloud

[0.1.3]

Added

  • Custom request headers via --header CLI flag and headers: config section
  • ${ENV_VAR} interpolation in header values for secret management
  • .env file loading at startup via dotenvy
  • SensitiveString type — secrets are redacted in debug output
  • {{ENV:VAR_NAME}} built-in template placeholder — resolved from environment at parse time
  • Load curve support — time-based VU scaling with stages: config
  • Threshold-gated CI — exit code 2 on threshold failure
  • Response template extraction — track specific fields from response bodies
  • OpenTelemetry tracing support via OTEL_EXPORTER_OTLP_ENDPOINT
  • JSON output mode (--output json) and file artifact (--output-file)
  • GitHub Actions workflows — CI, release (5-target cross-compiled binaries), publish, audit
  • Apache 2.0 license