Skip to content

CHANGELOG.MD

Latest commit

 

History

History
115 lines (66 loc) · 3.29 KB

CHANGELOG.MD

File metadata and controls

115 lines (66 loc) · 3.29 KB

Changelog

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

The format is based on Keep a Changelog.

[Unreleased]

Planned

  • Network baseline upgrade (sockets, RAM per connection, keep-alive).
  • New metrics and visualization improvements (queue wait times, service histograms).
  • Monte Carlo analysis with confidence intervals.

[0.1.1] – 2025-08-29

Added

  • Event Injection (runtime-ready):

    • Declarative events with start / end markers (server down/up, network spike start/end).
    • Runtime scheduler integrated with SimPy, applying events at the right simulation time.
    • Deterministic latency offset handling for network spikes (phase 1).

  • Improved Server Model:

    • Refined CPU + I/O handling with clearer queue accounting.
    • Ready queue length now explicitly updated on contention.
    • I/O queue metrics improved with better protection against mis-counting edge cases.
    • Enhanced readability and maintainability in endpoint step execution flow.

Documentation

  • Expanded examples on event injection in YAML.
  • Inline comments clarifying queue management logic.

Notes

  • This is still an alpha-series release, but now supports scenario-driven event injection and a more faithful server runtime model, paving the way for the upcoming network baseline upgrade.

[0.1.0a2] – 2025-08-17

Fixed

  • Quickstart YAML in README: corrected field to ensure a smooth first run for new users.

Notes

  • Minor docs polish only; no runtime changes.

[0.1.0a1] – 2025-08-17

Changed

  • Repository aligned with the PyPI 0.1.0a1 build.
  • Packaging metadata tidy-up in pyproject.toml.

CI

  • Main workflow now also triggers on push to main.

Notes

  • No functional/runtime changes.

[v0.1.0-alpha] – 2025-08-17

First public alpha of AsyncFlow — a SimPy-based, event-loop-aware simulator for async distributed systems.

Highlights

  • Event-loop model per server: explicit CPU (blocking), I/O waits (non-blocking), RAM residency.
  • Topology graph: generator → client → (LB, optional) → servers; multi-server via round-robin; stochastic network latency and optional dropouts.
  • Workload: stochastic traffic via simple RV configs (Poisson defaults).

Metrics & Analyzer

  • Event metrics: RqsClock (end-to-end latency).

  • Sampled metrics: ready_queue_len, event_loop_io_sleep, ram_in_use, edge_concurrent_connection.

  • Analyzer API (ResultsAnalyzer):

    • get_latency_stats(), get_throughput_series()
    • Plots: plot_latency_distribution(), plot_throughput()
    • Per-server: plot_single_server_ready_queue(), plot_single_server_io_queue(), plot_single_server_ram()
    • Compact dashboards.

Examples

  • YAML quickstart (single server).

  • Pythonic builder:

    • Single server.
    • Load balancer + two servers example with saved figures.

Tooling & CI

  • One-shot setup scripts (dev_setup, quality_check, run_tests, run_sys_tests) for Linux/macOS/Windows.
  • GitHub Actions: Ruff + MyPy + Pytest; system tests gate merges into main.

Compatibility

  • Python 3.12+ (Linux/macOS/Windows).
  • Install from PyPI: pip install asyncflow-sim.