FlagOS DevOps Central

This repository is the single source of truth for CI/CD logic across the FlagOS AI organization. It contains unified GitHub Actions (individual steps) and Reusable Workflows (entire pipelines) to ensure consistency, security, and speed across all our projects.

📁 Repository Structure

flagos-ai/FlagOps/
├── .github/
│   └── workflows/                  # Reusable Workflows & CI tests
│       ├── ci-shared.yml
│       ├── test-post-pytest-report.yml
│       └── test-post-benchmark-report.yml
├── actions/                        # Custom Actions (The "Individual Steps")
│   ├── setup-poetry/
│   │   └── action.yml
│   ├── notify-slack/
│   │   └── action.yml
│   ├── post-pytest-report/         # Upload pytest JSON reports
│   │   ├── action.yml
│   │   └── README.md
│   └── post-benchmark-report/      # Upload benchmark data with custom columns
│       ├── action.yml
│       └── README.md
└── README.md

---

🚀 How to use Shared Workflows

Reusable workflows allow you to standardize entire CI pipelines (e.g., testing, linting, and deploying) with just a few lines of code.

Usage Example

In your application repository (e.g., flagos-ai/model-api), create a file at .github/workflows/ci.yml:

name: CI

on:
  push:
    branches: [main]
  pull_request:

jobs:
  run-standard-ci:
    uses: flagos-ai/FlagOps/.github/workflows/ci-python.yml@main
    with:
      python-version: "3.11"
    secrets:
      HF_TOKEN: ${{ secrets.HF_TOKEN }}

---

🛠 How to use Shared Actions

Custom actions are discrete tasks used within your own existing jobs.

Usage Example

In any workflow file within the organization:

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      
      # Using a shared action from this repo
      - name: Setup Environment
        uses: flagos-ai/FlagOps/actions/setup-poetry@main
        with:
          install-dev-deps: "true"

Available Actions

Action	Description
setup-poetry	Set up Poetry environment
notify-slack	Send Slack notifications
post-pytest-report	Upload pytest JSON reports to backend
post-benchmark-report	Upload benchmark data with custom table columns

---

📌 Best Practices

1. Versioning

For production-critical repositories, avoid using @main. Instead, reference a specific version tag or commit SHA to prevent breaking changes:

• uses: flagos-ai/FlagOps/.github/workflows/ci-python.yml@v1.0.0
• uses: flagos-ai/FlagOps/actions/setup-poetry@v1.2.3

2. Contributions

1. Create a new branch for your changes.
2. Add your action or workflow following the folder structure above.
3. Document any required inputs or secrets in the local action.yml or workflow file.
4. Open a Pull Request for the DevOps team to review.

---

📞 Support

For questions regarding CI/CD standards or help with these actions, please contact the Infrastructure team or open an issue in this repository.