Configuration and Results
Use sample-config.json as a starting point. The JSON Schema defines all fields, defaults, and validation limits. Partial configurations are accepted; the server fills in defaults before validation and storage.
Every valid submission creates a new run with a unique run ID, including repeated configurations on the same project version. Earlier runs and episode results are retained, including failed and cancelled attempts. Callers decide whether to reuse an existing result or submit another experiment.
MariaDB stores runs in simulation_runs and episode measurements in
simulation_episodes.
For more information see the Developer page.