Edly Metrics

Modular, plugin-style metrics collection framework for engineering teams. Add Python collectors for different metric types (release success, lead time, PR throughput, etc.) and run them via a unified CLI. Supports data sources such as GitLab, GitHub, and Azure DevOps.

Quick start

1. Install dependencies

pip install -r requirements.txt

2. Set environment variables (example for GitLab)

export GITLAB_TOKEN=... 
export GITLAB_URL=https://git.overhang.io/api/v4

3. Run the CLI

python metrics_cli.py list
python metrics_cli.py run release_success --date-from 2024-01-01 --date-to 2024-12-31 --output ./files/release_success.csv

Project layout

  src/
  __init__.py
  base.py                # Collector interface
  registry.py            # Collector registry and discovery
  config.py              # Config models and loader
  datasources/
    __init__.py
    gitlab.py            # GitLab client wrapper
    github.py            # GitHub client stub
    azure_devops.py      # Azure DevOps client stub
  collectors/
    __init__.py
    release_success.py   # generates release success/failure rate
metrics_cli.py           # Typer-based CLI entry

Adding a new collector

1. Create a new file under src/collectors/your_collector.py implementing Collector.

2. Register it by calling registry.register_collector("your_name", YourCollector). If you follow the pattern inside collectors/__init__.py, import your module there so it self-registers on import.

3. Run it via CLI: python metrics_cli.py run your_name --output ./files/out.csv

Configuration

You can pass CLI flags directly or provide a YAML file with per-collector settings. Use gitlab_base_url for GitLab endpoint. Env vars (like tokens) are supported via python-dotenv and your shell environment.

Excel template

If you export to a .xlsx file, the framework will use pandas and openpyxl to write to Excel, which you can paste into your annual review template.