FlameIQ 1.0.0 Documentation

Deterministic, CI-native performance regression and evolution engine.

Make performance a first-class, enforceable engineering signal — without requiring any SaaS platform, cloud account, or vendor dependency.

pip install flameiq-core
flameiq init
flameiq baseline set --metrics benchmark.json
flameiq compare --metrics current.json --fail-on-regression

Note

FlameIQ OSS requires no internet connection, no account, and no API keys. It is fully offline and air-gap compatible.

---

Getting Started

• Installation
  • Requirements
  • Install from PyPI
  • Install in a virtual environment (recommended)
  • Install with optional dependencies
  • Install from source
  • Offline / air-gap installation
  • Upgrading
  • Uninstalling
• Quick Start
  • Step 1 — Initialise FlameIQ
  • Step 2 — Prepare benchmark output
  • Step 3 — Set a baseline
  • Step 4 — Compare on every pull request
  • Step 5 — Generate an HTML report
  • Step 6 — Advance the baseline
  • That’s it
• CI Integration
  • GitHub Actions
  • GitLab CI
  • Jenkins
  • CircleCI
  • Storing baselines
  • Exit codes

User Guides

• Configuration Reference
  • Complete example
  • Thresholds
  • Baseline strategies
  • Statistical mode
  • Noise handling
  • Config file location
• Baseline Strategies
  • Overview
  • last_successful
  • rolling_median
  • tagged
  • Switching strategies
• Building a Custom Provider
  • Provider contract
  • Minimal example
  • Register the provider
  • Use the provider
  • Testing your provider
  • Provider checklist

CLI Reference

• CLI Reference
  • Global options
  • flameiq init
  • flameiq run
  • flameiq compare
  • flameiq baseline
  • flameiq report
  • flameiq validate

Architecture

• Architecture Overview
  • Module layers
  • Dependency rules
  • Data flow
  • Design decisions
• Module Layer Reference
  • flameiq/schema/
  • flameiq/core/
  • flameiq/engine/
  • flameiq/storage/
  • flameiq/providers/
  • flameiq/reporting/
  • flameiq/cli/
• Schema Design
  • Versioning philosophy
  • Why dataclasses?
  • The flat() method

Specifications

• FlameIQ Performance Schema — Version 1
  • Overview
  • JSON structure
  • Field reference
  • Flat key format
  • Validation rules
  • Serialisation contract
  • Changelog
• Statistical Methodology Specification
  • Overview
  • Mann-Whitney U Test
  • Effect Size — Cohen’s d
  • Noise-resistant Median
  • Determinism guarantees
• Threshold Algorithm Specification
  • Overview
  • Step 1 — Compute change percent
  • Step 2 — Evaluate against threshold
  • Step 3 — Warning detection
  • Threshold format
  • Default threshold
  • Overall status
• Exit Code Specification
  • Using exit codes in CI
  • Using JSON output with exit codes

API Reference

• flameiq.schema — Data Models
  • Environment
  • LatencyMetrics
  • Metrics
  • SnapshotMetadata
  • PerformanceSnapshot
• flameiq.core — Comparison Engine
  • Comparator
  • Models
  • Thresholds
  • Errors
• flameiq.engine — Statistical Engine
  • Statistics
  • Baseline Strategies
• flameiq.storage — Local Storage
  • BaselineStore
• flameiq.providers — Plugin System
  • Base class
  • JSON Provider
  • pytest-benchmark Provider
  • Registry
• flameiq.reporting — HTML Reports
  • generate_report()

Contributing

• Development Setup
• RFC Process
• Testing Standards

Project

• Changelog
• Security Policy

Indices and tables

• Index

• Module Index

• Search Page