az-scout-plugin-latency-stats 2026.3.0

Inter-region latency statistics plugin for az-scout

az-scout-plugin-latency-stats

Inter-region latency statistics plugin for az-scout.

Features

• Latency dataset — static latency matrix based on Microsoft published statistics
• API endpoints — POST /plugins/latency-stats/matrix for pairwise latency matrices, GET /plugins/latency-stats/pairs to list all known pairs
• MCP tool — region_latency(source_region, target_region) returns indicative RTT between two Azure regions
• UI tab — interactive D3.js force-directed graph where regions are nodes and edges show latency in ms
• URL hash routing — #latency selects the plugin tab

Setup

pip install az-scout-plugin-latency-stats
az-scout  # plugin is auto-discovered

For development:

git clone https://github.com/lrivallain/az-scout-plugin-latency-stats
cd az-scout-plugin-latency-stats
uv sync --group dev
uv pip install -e .
az-scout  # plugin is auto-discovered

Structure

az-scout-plugin-latency-stats/
├── .github/
│   ├── copilot-instructions.md  # Copilot context for this plugin
│   └── workflows/
│       ├── ci.yml               # CI pipeline (lint + test, Python 3.11–3.13)
│       └── publish.yml          # Publish to PyPI on version tags
├── pyproject.toml
├── README.md
└── src/
    └── az_scout_latency_stats/
        ├── __init__.py          # Plugin class + module-level `plugin` instance
        ├── latency.py           # Static latency dataset + public API
        ├── routes.py            # FastAPI APIRouter (optional)
        ├── tools.py             # MCP tool functions (optional)
        └── static/
            ├── css/
            │   └── latency.css      # Plugin styles (auto-loaded via css_entry)
            ├── html/
            │   └── latency-tab.html # HTML fragment (fetched by JS at runtime)
            └── js/
                └── latency-tab.js   # Tab UI logic (auto-loaded via js_entry)

How it works

1. The plugin JS loads the HTML fragment into #plugin-tab-latency.
2. Regions are populated from the main app's regions global.
3. The user selects 2+ regions and clicks Show Latency Graph.
4. The plugin calls POST /plugins/latency-stats/matrix with the selected regions.
5. A D3.js force-directed graph renders regions as nodes with RTT-labelled edges.

API

# Pairwise matrix
curl -X POST http://localhost:8080/plugins/latency-stats/matrix \
  -H "Content-Type: application/json" \
  -d '{"regions": ["francecentral", "westeurope", "eastus"]}'

# All known pairs
curl http://localhost:8080/plugins/latency-stats/pairs

MCP tool

region_latency(source_region="francecentral", target_region="westeurope")

Quality checks

The scaffold includes GitHub Actions workflows in .github/workflows/:

• ci.yml — Runs lint (ruff + mypy) and tests (pytest) on Python 3.11–3.13, triggered on push/PR to main.
• publish.yml — Builds, creates a GitHub Release, and publishes to PyPI via trusted publishing (OIDC). Triggered on version tags (v*). Requires a pypi environment configured in your repo settings with OIDC trusted publishing.

Run the same checks locally:

uv run ruff check src/ tests/
uv run ruff format --check src/ tests/
uv run mypy src/
uv run pytest

To publish a release:

git tag v2026.2.0
git push origin v2026.2.0

Copilot support

The .github/copilot-instructions.md file provides context to GitHub Copilot about the plugin structure, conventions, and az-scout plugin API. It helps Copilot generate code that follows the project patterns.

Data source

Latency values are approximate median round-trip times from the Azure Network Latency page. Always validate with in-tenant measurements.

License

MIT

Disclaimer

This tool is not affiliated with Microsoft. All capacity, pricing, and latency information are indicative and not a guarantee of deployment success. Spot placement scores are probabilistic. Quota values and pricing are dynamic and may change between planning and actual deployment. Latency values are based on Microsoft published statistics and must be validated with in-tenant measurements.