I have read through all of Google's documentation on App Engine and various other blog posts but I still do not have a clear idea on the structure of a Django app running on App engine.
As I see things, App engine instances are attached to one SQL database and Google provides auto-scaling and auto load balancing that will replicate your SQL database according to current load.
What does this mean in terms of Django python code? All it seems to take to deploy a Django app is to run glcoud app deploy. If an App Engine instance's SQL database is scaled and replicated automatically in real time, is the Django app replicated along side them? Or will all HTTPS requests go through a single Django app no matter the replications?
In addition, is the Django backend code physically hugging the database it is connected to?
App Engine instances don’t come attached to anything. You need to also deploy a Cloud SQL instance either with MySQL or with Postgres and then use the Cloud SQL proxy to connect your application to the instance. It is not mandatory to use Cloud SQL, you can also run MySQL on Compute Engine or even on a third party platform but you will need to determine the proper way to connect it to your app.
As mentioned here You cannot autoscale a database instance, neither in Cloud SQL nor in any other infrastructure support. What Google Cloud SQL provides you is with a fully managed instance, this means that you don't have to worry about backups, patches, and fixes. Also, Cloud SQL provides you with High availability in the case of your instance fails to reduce a lot the downtime of your application, also provides you with an easy way to add read replicas, increasing the performance of your reads operation.
What does this mean in terms of Django python code? Probably you want to take a look at the Github sample and the running Django on App Engine quickstart.
All it seems to take to deploy a Django app is to run glcoud app deploy. That's not precise, you can run the glcoud app deploy to deploy ANY valid application to App Engine, in python a minimal application requires a main.py file, a requirements.txt, and an app.yaml doc. For Django you will need your valid Django application + your app.yaml file (look at git repo in point 3)
If an App Engine instance's SQL database is scaled and replicated automatically in real-time, is the Django app replicated alongside them? Again there is not such an App Engine SQL database. The database instance will run independently of your application in App Engine. When your App Engine application auto-scales your Cloud SQL instance (or compute engine if you decide to go that way) will remain exactly as it was before, it is "agnostic" of App Engine and its behavior and the same for App Engine, your App Engine instances are agnostic of the replication behavior of your Cloud SQL. You will have just a connection between them sending requests back and forth.
Or will all HTTPS requests go through a single Django app no matter the replications? You can have multiple instances in App Engine running your application at the same time, they will send requests back and forth your Cloud SQL database via the Cloud SQL proxy, not HTTP request.
In addition, is the Django backend code physically hugging the database it is connected to? Again, Your Django/App Engine App is one thing and your Cloud SQL database is another different thing, they will talk with each other via the Cloud SQL proxy. So no hugging or physical contact, just "social distancing"
Hope this helps