source:
  type: datahub-gc
  config:
    # Whether to run the recipe in dry-run mode or not
    dry_run: false
    # Cleanup expired tokens
    cleanup_expired_tokens: true
    # Whether to truncate elasticsearch indices or not which can be safely truncated
    truncate_indices: true

    # Cleanup DataProcess Instances
    dataprocess_cleanup:
      retention_days: 10
      # Delete empty Data Jobs (if no DataProcessInstance associated with the DataJob)
      delete_empty_data_jobs: true
      # Delete empty Data Flow (if no DataJob associated with the DataFlow)
      delete_empty_data_flows: true
      # Whether to hard delete entities or soft delete them
      hard_delete_entities: false
      # Keep the last n dataprocess instances
      keep_last_n: 5
    soft_deleted_entities_cleanup:
      # Delete soft deleted entities which were deleted 10 days ago
      retention_days: 10
    execution_request_cleanup:
      # Minimum number of execution requests to keep, per ingestion source
      keep_history_min_count: 10
      # Maximum number of execution requests to keep, per ingestion source
      keep_history_max_count: 1000
      # Maximum number of days to keep execution requests for, per ingestion source
      keep_history_max_days: 30
      # Number of records per read operation
      batch_read_size: 100
      # Global switch for this cleanup task 
      enabled: true