I'd be cool to use your approach to build a "frontend" for monitoring Python's TQDM output. Especially useful for multiprocess applications with scores of independent, long running processes like data ingestion pipelines. TQDM does support monitoring multiple processes but is not flexible enough for all use cases.
