zylon-ai/private-gpt · error · HTTPException

Async ingestion is not supported by the configured scheduler

Error message

Async ingestion is not supported by the configured scheduler

What it means

Error "Async ingestion is not supported by the configured scheduler" thrown in zylon-ai/private-gpt.

Source

Thrown at private_gpt/server/ingest/ingest_router.py:679

    * PDFs generate one document per page
    * Large files are automatically split into chunks
    * Progress can be monitored via /artifacts/ingest/async/{task_id}
    * Resulting context is available in chat/completion APIs
    * Ingested content can be filtered using metadata

    Important to know:
    - Since binary data cannot be passed directly to Celery tasks, files are
      first uploaded to a temporary S3 bucket. The Celery task then retrieves
      the file from S3 for processing. This can incur additional time.

    """
    scheduler = request.state.injector.get(IngestionSchedulerFactory).get()
    try:
        task_id = scheduler.ingest_async(body)
    except NotImplementedError as exc:
        from fastapi import HTTPException

        raise HTTPException(
            status_code=501,
            detail="Async ingestion is not supported by the configured scheduler",
        ) from exc
    return Task(task_id=task_id)


def ingest_data_sync(
    collection: str,
    artifact: str,
    data_path: Path,
    service: IngestService,
    metadata: dict[str, Any] | None,
) -> IngestResponse:
    """Internal helper function for synchronous document ingestion processing.

    Handles the core ingestion workflow including artifact initialization
    and vector index population for both file and text ingestion.
    """

View on GitHub (pinned to 4a030776a3)

Solutions

1. Configure a scheduler that supports async ingestion (e.g. set scheduler.chat.mode='arq' with the arq worker running).
2. Use the synchronous ingestion endpoint instead of async ingestion.
3. Enable and configure the required task queue backend (Redis/arq) so the scheduler supports async ingestion.

When it happens

Trigger: Raised when an async ingestion request is made but the configured ingestion scheduler only supports synchronous execution.

Common situations: Sending an async ingest request while running the default local scheduler; switch to a scheduler that supports async jobs or use the synchronous endpoint.

AI-assisted analysis of zylon-ai/private-gpt@4a030776a3 (2026-08-15). Data as JSON: /api/errors/026a6865b6ab8eee. Report an issue: GitHub.