Django Tenant Schemas Q is a custom implementation for Django-Q cluster for projects that make use Django Tenant Schemas package to achieve multi-tenancy in their Django projects with Postgres as database service. Using this package, developer can setup a Django project to be multi-tenant and setup Django-Q cluster to work per tenant/schema basis. This package works with any other broker apart from Django ORM.
Although there is package tenant-schemas-celery supporting celery for multi-tenancy, I was quite intrigued by the django-q project and was exploring the possibility of making django-q work with django-tenant-schemas.
pip install django-tenant-schemas-q
This should install django-tenant-schemas and django-q.
As a developer I often find that setting up projects, packages and getting configurations right is sometimes a relatively tough job. In order to make this work nicely, just follow the steps below.
-
Refer to settings required by django_tenant_schemas mentioned here
-
Setup django_tenant_schemas as per the instructions in the above mentioned link. This will take care of setting up of django_tenant_schemas.
-
Add django_q to TENANT_APPS setting. This will setup models like Task, Schedule on every tenant post migrations.
-
Follow it up with setting up the Django-Q cluster that needs to be run in order to process tasks and run our scheduler.
-
While setting up configuration for cluster, make use of any other broker supported by Django-Q apart from Django ORM.
-
Finally add a setting SCHEMAS_TO_BE_EXCLUDED_BY_SCHEDULER in your settings file. The value for this setting is a list of schema that you wish to exclude from scheduler, If not specified, scheduler will exclude a schema by name public by default.
Assuming you have followed the instructions perfectly, you should now have settings looking like
SHARED_APPS = ['tenant_schemas', 'app_containing_tenant_model ', 'any_other_app_needed']
TENANT_APPS = ['django_q', 'standard_django_apps', 'any_other_app_needed']
INSTALLED_APPS = ['tenant_schemas', 'app_containing_tenant_model', 'django_q', 'standard_django_apps', 'any_other_app_needed']
TENANT_MODEL = 'app_name.ModelName'
SCHEMAS_TO_BE_EXCLUDED_BY_SCHEDULER = ['public']
Q_CLUSTER = {} # Configuration for Q Cluster
What these settings will do is simple. It will simply create Task & Schedule models per schema. Django-Q uses these modules to store the tasks and schedules. With these models now setup on per schema/tenant basis the things become a bit cleaner.
To run the cluster, use the command python manage.py mscluster
Once the command is fired, the cluster will start and accept the tasks and schedules.
Now, given that are schedules are tasks that are executed as per the frequency having them per schema/tenant basis is something that I wanted to achieve. This would give me ability to configure same tasks with different times and the system will work as expected.
To allow adding a task or schedule with database schema awareness, package contains utilities to make it is seamless.
Add the line from django_tenant_schemas_q.utils import QUtilities to the line from where you wish to add a task or a schedule
To add an async task you can use
QUtilities.add_async_task(func_name_as_string, *args, **kwargs)
You can add sync=True in kwargs to run the task synchronously.
To create a new schedule you can use
QUtilities.create_schedule(func_name_as_string, *args, **kwargs)
To get a result of a single task
QUtilities.get_result(task_id, wait=0, cached=Conf.CACHED)
To get a result of a group
QUtilities.get_result_group(group_id, failures=False, wait=0, count=None, cached=Conf.CACHED)
To fetch a single task
QUtilities.fetch_task(task_id, wait=0, cached=Conf.CACHED)
To fetch a group of tasks
QUtilities.fetch_task_group(group_id, failures=True, wait=0, count=None, cached=Conf.CACHED)
To get count of groups
QUtilities.get_group_count(group_id, failures=False, cached=Conf.CACHED)
To delete a group of tasks
QUtilities.delete_task_group(group_id, tasks=False, cached=Conf.CACHED)
To delete a task from cache
QUtilities.delete_task_from_cache(task_id, broker=None)
To get the size of the queue
QUtilities.get_queue_size(broker=None)
To add multiple async tasks using Iter
QUtilities.add_async_tasks_from_iter(func, args_iter, **kwargs)
For this use Iter from django_tenant_schemas_q.custom module.
To create a chain of tasks using Chain
QUtilities.create_async_tasks_chain(chain, group=None, cached=Conf.CACHED, sync=Conf.SYNC, broker=None)
For this use Chain from django_tenant_schemas_q.custom module.
There is a test django project in the repository.
- Clone the repository
- Run
docker-compose -f test-compose.yml build - Run
docker-compose -f test-compose.yml up -d - Run
docker-compose -f test-compose.yml run backend python manage.py setupdata - Run
docker-compose -f test-compose.yml run backend python manage.py test --keepdb
Full credit to authors https://github.com/Koed00 of Django-Q and https://github.com/bernardopires of Django-Tenant-Schemas for two wonderful packages.