๐ฑ Experimental environment to develop on energy- and carbon-aware cloud-native software.
- Local cloud infrastructure
- Energy consumption monitoring stack
- Carbon emission monitoring stack
- Machine learning infrastructure and workloads
- Visualization stack
- Execution management tool
This project uses the just command runner for ergonomic development. To install with cargo:
cargo install just
Download and install all necessary requirements:
just install requirements
To start the cluster and deploy all components:
Note: Deployment might fail if a VPN is active.
just init
To deploy specific components:
just deploy infra # spin up a local cluster, without deploying other components
just deploy [energy-monitoring|carbon-monitoring|machine-learning|visualization]
To run a workload, for example to test energy consumption or carbon emission monitoring:
just run [machine-learning] # run all workloads defined in machine-learning
If deployed, visualizations can be accessed with:
just view # opens all dashboards
just view [kepler|flyte]
To remove everything:
just teardown # remove the entire cluster, including all components
To remove specific components:
just teardown [energy-monitoring|carbon-monitoring|machine-learning|visualization]