How to use Django with Celery+Reds within DO App Platform?

Question

On my local machine, I have a very simple proof of concept that uses docker-compose to run Django in one container, Redis in a second container, and a Celery worker in a third container. It “works on my machine”.

How do I deploy this to DigitalOcean? How do I implement blue/green deployment using DigitalOcean assets?

For a Django application, without Celery and Redis, it seems straightforward. I can build, tag, and push the Django container into the DO registry, and then use doctl to deploy my application, update environment variables, and switch the floating IP between the green and blue servers. Great.

docker build -t <container> .
docker tag <container>:latest registry.digitalocean.com/<repo>/<container>:latest
docker push registry.digitalocean.com/<repo>/<container>:latest
doctl apps create-deployment <app-id>
doctl apps update <app-id> --spec <prod.yaml>
doctl compute floating-ip-action assign <floating-ip> <new-prod-droplet-id>

Now how do I set up Celery and Redis? I know I can stand up a Redis DB cluster, but do I need a separate droplet to run the Celery worker? Or is there a way to run the Celery worker from the same compute resource as my containerized Django app? What’s the best practice here?

Bobby · Answer

Hi there,

This seems like a great setup. I would personally go for a managed Redis Cluster just so that you don’t have to manage it yourself on a separate Droplet.

Regarding the Celery service, as this is just a POC, you could deploy the Celery container on the same Droplet as your Django app.

If you are not deploying on a Droplet but the App Platform itself, then you could add the Celery service as a worker:

https://docs.digitalocean.com/products/app-platform/how-to/manage-workers/

To do that, you would just have to update your prod.yaml file and include the Celery service as a component to your Django App.

Hope that this helps!

Best,

Bobby