zylon-ai/private-gpt · error · ValueError

The arq worker mode does not support arguments yet

Error message

The arq worker mode does not support arguments yet

What it means

Raised by run_arq when the 'arq' worker mode is invoked with extra command-line arguments. Unlike the celery mode, which forwards args to the celery CLI, the arq launcher ignores argv entirely and refuses to run rather than silently dropping user-specified options like a queue name or concurrency flag.

Source

Thrown at private_gpt/worker/modes.py:125

        "0.0.0.0",
        "--port",
        os.environ.get("API_PORT", "8090"),
        "--no-access-log",
        "--log-level",
        "critical",
    ]


def _with_healthcheck(command: list[str]) -> list[list[str]]:
    commands = [command]
    if os.environ.get("API_ENABLED", "true").lower() == "true":
        commands.append(_healthcheck_command())
    return commands


def run_arq(args: Sequence[str]) -> None:
    if args:
        raise ValueError("The arq worker mode does not support arguments yet")
    arq_app = importlib.import_module(f"{_app_module()}.arq")
    arq_app.run_arq_worker()


def run_celery(
    args: Sequence[str],
    *,
    celery_settings_provider: Callable[[], CelerySettings] = lambda: settings().celery,
) -> None:
    celery_args = _build_celery_args(celery_settings_provider())
    typer.echo(f"Starting celery worker with args: {' '.join(celery_args)}")
    command = [
        sys.executable,
        "-m",
        "celery",
        "--app",
        f"{_app_module()}.celery",
        "worker",

View on GitHub (pinned to 4a030776a3)

Solutions

1. Remove all arguments from the arq worker invocation - configure it via settings/env instead (e.g. celery/arq settings blocks in settings.yaml)
2. If you need celery-style CLI args, switch the worker mode to celery, which builds and forwards them
3. Pass configuration through PGPT_* environment variables or the settings file rather than argv

Example fix

# before
PGPT_WORKER_MODE=arq private-gpt worker arq --concurrency 4

# after
PGPT_WORKER_MODE=arq private-gpt worker arq
# tune concurrency in settings.yaml instead

import os, sys

mode = os.environ.get("PGPT_WORKER_MODE", "")
if mode == "arq" and len(sys.argv) > 1:
    raise SystemExit("arq mode takes no CLI args; configure via settings.yaml")

null

try:
    run_arq(args)
except ValueError as e:
    if "does not support arguments" in str(e):
        args = ()  # retry without args, moving config to settings
    else:
        raise

When it happens

Trigger: Running the worker CLI as `private-gpt worker arq --queue default` or `... arq -w 4`; a deployment script or Docker CMD that appends generic worker flags to every worker mode; a supervisor/systemd unit reused from the celery configuration and pointed at arq.

Common situations: Migrating a deployment from celery to arq mode and keeping the old arg flags; wrapper scripts that pass '--concurrency' style options unconditionally; container orchestrators templating the same args array into all worker containers.

Related errors

• Binary not found in PATH: {binary!r}
• Server did not become ready within {_HEALTH_TIMEOUT}s
• Server is already running with PID {existing_pid}
• Unknown command: {cmd!r}
• Unsupported scheduler.chat.mode={self.scheduler.chat.mode!r}

AI-assisted analysis of zylon-ai/private-gpt@4a030776a3 (2026-08-15). Data as JSON: /api/errors/98aff16e61e29606. Report an issue: GitHub.