django-tasks-db-async 0.0.1

Async worker for django-tasks-db with OpenTelemetry instrumentation

django-tasks-db-async

Async worker for django-tasks-db with OpenTelemetry instrumentation.

---

Table of Contents

• Installation
• Configuration
• Running the worker
• OpenTelemetry
• License

Installation

pip install django-tasks-db-async

Configuration

Add both django_tasks_db and django_tasks_db_async to INSTALLED_APPS, then configure the TASKS setting to use django_tasks_db.DatabaseBackend:

INSTALLED_APPS = [
    ...
    "django_tasks_db",
    "django_tasks_db_async",
]

TASKS = {
    "default": {
        "BACKEND": "django_tasks_db.DatabaseBackend",
        "QUEUES": ["default"],
    },
}

Run the database migrations to create the task result table:

python manage.py migrate

Running the worker

Start the worker with the db_async_worker management command:

python manage.py db_async_worker

Options

Option	Default	Description
--queue-name	default	Queue(s) to process. Separate multiple with a comma. Use * to process all queues.
--interval	1	Seconds to wait when the queue is empty before polling again.
--concurrency	10	Number of concurrent worker coroutines.
--batch	False	Process all outstanding tasks and exit.
--max-tasks	unlimited	Maximum number of tasks each worker will execute before stopping.
--backend	default	Backend alias to operate on (as configured in TASKS).
--worker-id	auto	Worker identifier, must be unique across the pool.

[!NOTE] Each worker coroutine holds one database connection for the duration of its polling loop. A --concurrency of 10 (the default) can therefore open up to 10 simultaneous connections. Make sure your database connection pool (or CONN_MAX_AGE) is sized accordingly, and consider lowering --concurrency if your database has a low connection limit.

Examples

Process tasks from all queues with 4 concurrent workers:

python manage.py db_async_worker --queue-name='*' --concurrency=4

Process tasks from specific queues:

python manage.py db_async_worker --queue-name=default,emails,notifications

Run in batch mode — process outstanding tasks and exit (useful for cron jobs):

python manage.py db_async_worker --batch

Signals

The worker dispatches the standard task_started and task_finished Django signals around each task execution, compatible with both django.tasks (Django 6.0+) and the standalone django-tasks package.

OpenTelemetry

The worker records a span for each task execution using the OpenTelemetry messaging semantic conventions. No additional configuration is required — the worker picks up whatever TracerProvider is active in the process.

Each span includes the following attributes:

Attribute	Value
messaging.system	django_tasks_db
messaging.operation.type	process
messaging.destination.name	queue name
messaging.message.id	task result UUID
messaging.client.id	worker ID
messaging.operation.name	fully qualified task function name

Failed tasks set the span status to ERROR and record the exception as a span event.

License

django-tasks-db-async is distributed under the terms of the MIT license.