Troubleshooting¶

Worker Starts but Jobs Never Run¶

Check:

1. Queue name matches task queue name.
2. Task module is imported (or configured in settings.tasks for autodiscovery).
3. Backend in producer and worker is the same environment/config.

Useful commands:

asyncmq info backend
asyncmq queue list
asyncmq job list --queue <queue> --state waiting

Jobs Stay in waiting¶

Common causes: - queue paused (asyncmq queue info <queue>) - no worker consuming that queue - unresolved dependencies (depends_on parents not completed)

Delayed Jobs Never Fire¶

Check scan_interval and worker process health. Delayed jobs are moved by delayed_job_scanner, so a worker loop must be running.

Stalled Jobs Not Recovered¶

enable_stalled_check=True only enables heartbeat recording. You also need a recovery scheduler loop:

from asyncmq.core.stalled import stalled_recovery_scheduler
await stalled_recovery_scheduler()

Postgres Backend Errors on Missing Tables¶

Initialize schema before running workers/producers:

import anyio
from asyncmq.core.utils.postgres import install_or_drop_postgres_backend

anyio.run(install_or_drop_postgres_backend)

Dashboard Redirects to Login Unexpectedly¶

Check:

• AsyncMQAdmin(enable_login=True, backend=...) has a valid backend
• session middleware is enabled (include_session=True or host app equivalent)
• mount path/prefix is consistent with configured dashboard prefix

CLI Uses Wrong Backend¶

Ensure the same settings module is exported in your shell/session:

export ASYNCMQ_SETTINGS_MODULE=myapp.settings.AppSettings
asyncmq info backend