Django Transaction Task Queue
=============================
A short and simple Celery replacement for my Django projects.
* *Database is the only backend.* The task is a simple Django model, it uses the same transaction and connection as other models. No more ``transaction.on_commit`` hooks to schedule tasks.
* *Tasks do not produce any results* and there is no "result backend". If you need to store results, pass a unique key into the task and store the result in some DIY model.
* *You can not wait on the response from the task*, there is no ``get`` or ``join`` to get the task result. I think a thread pool or process pool is better for those requirements.
* *No workflows and chaining of jobs.*
* *ETA (estimated time of arrival) is a first-class citizen.* It does not depend on whether backends support the feature or not.
* *The worker is a single-threaded process*, you start several of those to scale. I have seen too many issues with autoscaling workers, worker processes killed by OS, workers stuck: simple is better.
* *No prefetching or visibility timeouts*. The worker picks the first available task and processes it.
* *Dead letter queue* built in. You get access to failed tasks and can retry them.
* *Django admin for monitoring.* You can view pending tasks, failed, and "dirty" (crashed in the middle of work). Failed and "dirty" tasks can be retried from the same Django admin.
* *Easy to get the metrics* from Django shell and export to your favorite monitoring tool
* *Task records are removed after successful execution*. Unlike Celery SQLAlchemy's backend, records are removed so you don't have to care about archiving. It also keeps the table small, properly indexed and efficient.
Raw data
{
"_id": null,
"home_page": "https://github.com/aivarsk/django-transactional-task-queue",
"name": "django-transactional-task-queue",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.10",
"maintainer_email": null,
"keywords": null,
"author": "Aivars Kalvans",
"author_email": "aivars.kalvans@gmail.com",
"download_url": "https://files.pythonhosted.org/packages/4b/5b/8ca743a0df2149cc2e26f312f3ab4887a78dfd0a20c6d4fd37f21fd50e36/django_transactional_task_queue-0.1.4.tar.gz",
"platform": null,
"description": "Django Transaction Task Queue\n=============================\n\nA short and simple Celery replacement for my Django projects.\n\n* *Database is the only backend.* The task is a simple Django model, it uses the same transaction and connection as other models. No more ``transaction.on_commit`` hooks to schedule tasks.\n* *Tasks do not produce any results* and there is no \"result backend\". If you need to store results, pass a unique key into the task and store the result in some DIY model.\n* *You can not wait on the response from the task*, there is no ``get`` or ``join`` to get the task result. I think a thread pool or process pool is better for those requirements.\n* *No workflows and chaining of jobs.*\n* *ETA (estimated time of arrival) is a first-class citizen.* It does not depend on whether backends support the feature or not.\n* *The worker is a single-threaded process*, you start several of those to scale. I have seen too many issues with autoscaling workers, worker processes killed by OS, workers stuck: simple is better.\n* *No prefetching or visibility timeouts*. The worker picks the first available task and processes it.\n* *Dead letter queue* built in. You get access to failed tasks and can retry them.\n* *Django admin for monitoring.* You can view pending tasks, failed, and \"dirty\" (crashed in the middle of work). Failed and \"dirty\" tasks can be retried from the same Django admin. \n* *Easy to get the metrics* from Django shell and export to your favorite monitoring tool\n* *Task records are removed after successful execution*. Unlike Celery SQLAlchemy's backend, records are removed so you don't have to care about archiving. It also keeps the table small, properly indexed and efficient.\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "A simple Celery-like transactional task queue for Django",
"version": "0.1.4",
"project_urls": {
"Homepage": "https://github.com/aivarsk/django-transactional-task-queue"
},
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "4b5b8ca743a0df2149cc2e26f312f3ab4887a78dfd0a20c6d4fd37f21fd50e36",
"md5": "417b47e14b33781c508e51e149f109f1",
"sha256": "6d4ef35871d3194c2dc49f877d0db067550880e0fb421f3691e5f1bea4180e7d"
},
"downloads": -1,
"filename": "django_transactional_task_queue-0.1.4.tar.gz",
"has_sig": false,
"md5_digest": "417b47e14b33781c508e51e149f109f1",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.10",
"size": 8301,
"upload_time": "2024-07-23T14:28:22",
"upload_time_iso_8601": "2024-07-23T14:28:22.293808Z",
"url": "https://files.pythonhosted.org/packages/4b/5b/8ca743a0df2149cc2e26f312f3ab4887a78dfd0a20c6d4fd37f21fd50e36/django_transactional_task_queue-0.1.4.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-07-23 14:28:22",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "aivarsk",
"github_project": "django-transactional-task-queue",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "django-transactional-task-queue"
}
JSON
