Scheduling Custom Django Management Commands
Last updated October 18, 2021
Table of Contents
Running scheduled tasks that rely on data or code in your Django application is simple with custom django-admin commands and Heroku Scheduler.
Create a custom management command
Create the Python file
Custom Django management commands are nested within a Django project’s apps. For this tutorial, if the Django project contains more than one app, decide which app the custom management command will most closely relate to, and use that app to follow the steps below.
A Django project contains one or more apps. In this tutorial, you’ll be working within one of the apps and not the overall Django project.
The Python file that contains the code for the custom command is nested inside several directories:
- At the root level of the Django app, create a directory called
management
- Within
management
create acommands
directory - Within
commands
create a file named something similar toprint_book_titles.py
. Stylistically, the name you choose for this file should reference the outcome of the task when it runs.
At this stage, the app’s file tree will look similar to this:
some_app/
__init__.py
models.py
management/
commands/
print_book_titles.py
tests.py
views.py
Django automatically maps the name of the command file to the command that is run in the CLI (python manage.py print_book_titles
)
Code the command’s logic
The following code snippet is for a short custom management command. View the Django documentation on Writing custom django-admin commands for guidance on more complex needs.
from django.core.management.base import BaseCommand, CommandError
from some_app.models import Book
class Command(BaseCommand):
help = 'Prints all book titles in the database'
def handle(self):
try:
books = Book.objects.all()
for book in books:
self.stdout.write(self.style.SUCCESS(book.title))
except FieldDoesNotExist:
self.stdout.write(self.style.ERROR('Field "title" does not exist.'))
return
self.stdout.write(self.style.SUCCESS('Successfully printed all Book titles'))
return
Include a return
statement at the end of each logical flow of the handle()
function so Heroku Scheduler knows when it can shut down. This ensures you are only charged for dyno time that is needed to run the job.
Verify the command works as expected
Ensure the command works by running it in a one-off dyno:
- From the CLI, run
heroku run bash -a your-app-name
- Navigate to the directory containing
manage.py
- Run
python manage.py <your_custom_command>
Setup Heroku Scheduler
Provision the Add-on
If the Heroku Scheduler Add-On is not already provisioned on the Heroku app, follow the steps in Heroku Scheduler - Installing the Add-On now.
Schedule the job
Open the Heroku Scheduler tool by following the steps in Heroku Scheduler - Scheduling Jobs. When prompted, enter python manage.py <your_custom_command>
as the command to run.
If the manage.py
file is not at the root level of the Heroku app’s git repo, the command that is run within Heroku Scheduler will need to first navigate to the directory that contains the manage.py
file. The command would look something like cd my-project && python manage.py <your-command-name>
.