Architecture

EcoGuard is built as a pipeline of data collection, analysis, storage, and presentation.

System layers

1. External data sources
   • GitLab API for pipeline and job metadata
   • Electricity Maps for live carbon intensity
2. Processing layer
   • collect_real_data.py
   • agent modules in agents
3. Storage and serving
   • JSON metrics in dashboards/data
   • Flask API in api_server.py
4. Presentation layer
   • static dashboard files in public
5. Automation layer
   • GitHub Actions and GitLab CI/CD workflows

Main data flow

1. A commit, merge request, or scheduled job starts the pipeline.
2. The collector fetches GitLab job data and external carbon intensity data.
3. The agents convert runtime into energy and CO₂ estimates.
4. The dashboard data agent writes JSON summaries.
5. The backend or static site reads those JSON files.
6. The dashboard displays trends, recommendations, and goals.

Important files

• collect_real_data.py
• api_server.py
• agents/carbon_footprint.py
• agents/resource_optimization.py
• agents/eco_friendly_deployment.py
• agents/dashboard_data.py