Back to docs
Troubleshooting

Troubleshooting

Local mode issues

  • ModuleNotFoundError: ensure uv sync has installed the package and run via uv run.
  • Task timeout: increase timeout_seconds on the task or optimize the task.
  • Dependency cycle detected: check for circular dependencies in the DAG.
  • Artifact serialization failed: ensure task outputs are JSON-serializable or return a LocalArtifact.

Dependency package issues

  • Validation errors on Save & Build: Two packages may require mutually exclusive versions of a shared transitive dependency. Try relaxing one of the conflicting version constraints, or check that all packages are available for your selected Python version.
  • Duplicate package names: The platform normalizes names using PEP 503 (lowercase, hyphens only). Variations like scikit-learn, scikit_learn, and Scikit.Learn are treated as the same package and deduplicated automatically.
  • Version conflicts during bulk import: When importing a requirements.txt that contains a package already in the list with a different version, use the amber conflict banner to choose between the existing and incoming version for each conflict.
  • Build creates a duplicate record: This should not happen. The platform tracks the package ID from the initial creation, so clicking Save & Build again updates the existing record. If you see duplicates, check that you are editing the same package and not creating a new one.

Where to look

  • Run logs: ~/.dagy/runs/<run_id>/run.log
  • Task logs: ~/.dagy/runs/<run_id>/task_runs/<task_id>/task.log
  • Artifacts: ~/.dagy/artifacts/<run_id>/<task_id>/