Skip to content

CHANGELOG.md

Latest commit

 

History

History
115 lines (74 loc) · 4.63 KB

CHANGELOG.md

File metadata and controls

115 lines (74 loc) · 4.63 KB
Raw

Changelog

All notable changes to apiforgepy are documented here.

Format: Keep a Changelog — versioning follows Semantic Versioning.

[Unreleased]

[4.0.0] — 2026-07-08

Breaking Changes

  • Cloud mode now sends the X-APIForge-Key header instead of X-API-Key, matching the saas-api rename (CDC §8.4.1). Cloud ingest against the current APIForge API requires this version — 3.x and earlier send the old header and are rejected with 401. The header is internal to the SDK, so no code change is needed on your side beyond upgrading.

Changed

  • The local dashboard now serves its React/Babel runtime from files vendored inside the package instead of downloading them from a CDN (jsDelivr) on first launch. The dashboard works fully offline and makes no third-party requests — privacy-first parity with the Node and PHP SDKs. Bundled versions: react 18.3.1, react-dom 18.3.1, @babel/standalone 7.29.7.

Migration guide

pip install "apiforgepy>=4.0.0"

No configuration change is required — the header is set internally by CloudTransport. Upgrade any service running the SDK in cloud mode.

[3.0.0] — 2026-06-04

Breaking Changes

  • flush_interval parameter removed from ApiForgeMiddleware — passing it now raises TypeError. The flush window is fixed at 60 seconds.
  • env no longer falls back to os.environ.get("ENV") — must be passed explicitly. Default is now 'production'.
  • release no longer falls back to os.environ.get("APP_VERSION") — must be passed explicitly. Default is now None.

Added

  • bytes_avg field: average response body size (bytes) per route per bucket, sourced from the Content-Length response header
  • inflight_avg and inflight_max per route — inflight concurrency count captured via ASGI scope and aggregated per minute bucket

Migration guide

# Before (v2.x)
app.add_middleware(
    ApiForgeMiddleware,
    flush_interval=30_000,                        # ← remove (TypeError in v3)
    env=os.environ.get("ENV", "production"),       # ← pass explicitly
    release=os.environ.get("APP_VERSION"),         # ← still OK (your app reads the env var)
)

# After (v3.0)
app.add_middleware(
    ApiForgeMiddleware,
    env="production",    # set explicitly
    release="v1.4.0",    # set explicitly
)

[1.0.3] — 2026-05-15

Added

  • DRIFT insight type: detects progressive latency degradation using ordinary least squares over the last 30 days — emitted when slope ≥ 5ms/day over 7+ data points, with a 30-day projection
  • DRIFT filter chip added to the dashboard Insights view
  • 60 unit tests covering aggregator, database, insights and middleware

Fixed

  • CI badge was pointing to main with no workflow run — CI now triggers on both main and dev

1.0.0 — 2026-05-15

Fixed

  • /api/summary now returns a flat response structure (calls_24h, error_rate_24h, avg_p90_24h, active_routes) matching the React dashboard frontend — fixes "0 requests" display

Added

  • /api/global-timeseries endpoint consumed by the dashboard overview chart
  • /api/releases endpoint (returns [] for now — release tracking planned)

Changed

  • Dashboard UI now loads React and Babel from jsDelivr CDN instead of local assets, making ui.html SDK-agnostic

0.1.0 — 2026-05-15

Added

  • Starlette/FastAPI middleware ApiForgeMiddleware — drop-in observability with zero mandatory configuration
  • Local-first mode with SQLite storage via Python's built-in sqlite3 module (requires Python ≥ 3.11)
  • Per-endpoint metrics: P50 / P90 / P99 latency, request count, 2xx / 4xx / 5xx breakdown
  • In-memory aggregation with configurable flush interval (default: 60s), thread-safe
  • Circuit breaker on the transport layer — middleware never crashes the host application
  • Built-in dashboard on port 4242 with dark theme, Chart.js time series, routes table and insights panel
  • REST API: /api/summary, /api/routes, /api/timeseries, /api/insights
  • Three automatic insight types: ANOMALY (Z-score P99), DEAD (endpoint inactive 21+ days), PERF/OK (regression or improvement after a release)
  • API Health Score (0–100) combining availability, performance, stability and quality
  • Configurable sampling rate, ignored paths, environment label, release tag and service name
  • middleware.shutdown() for graceful teardown (flushes buffer, closes SQLite)