How to create a secure infrastructure for complex operations?

Question

Django

The question is more theoretical - I'm looking for useful materials.

I want to find a good example (any technology stack) when services are created with large complex asynchronous tasks. A purely hypothetical example - there is a service that takes a project from git and does something with it and, say, uploads it somewhere. The process is quite long and complicated. Users register on the service and can create such processes like cron. Using the example of Django + Celery (RQ), creating a queue where tasks scheduled from users will be added and executed is still clear. WHAT NEXT? I want the task to start in some isolated environment and be executed and the environment to die. So that the tasks cannot somehow harm each other. Here's how to properly stir up this environment for an asynchronous task that will be raised from Celery (or the same HangFire from .NET for example). For example, I did something similar on Heroku and simply raised a separate dyno with bash and ran a script there that passed parameters for the base. the script worked and dyno died. Beautiful, but Heroku. On the Google cloud, you can probably stir up something like this on Cloud Functions, but this is the Google Platform. And what are the solutions for a regular VPS on Linux? Docker? Maybe some other elegant solutions to raise such micro-isolated environments for long asynchronous tasks? In order not to bother (or minimally bother) with the allocation / control of resources, control the number of simultaneously running scripts / environments. And what are the solutions for a regular VPS on Linux? Docker? Maybe some other elegant solutions to raise such micro-isolated environments for long asynchronous tasks? In order not to bother (or minimally bother) with the allocation / control of resources, control the number of simultaneously running scripts / environments. And what are the solutions for a regular VPS on Linux? Docker? Maybe some other elegant solutions to raise such micro-isolated environments for long asynchronous tasks? In order not to bother (or minimally bother) with the allocation / control of resources, control the number of simultaneously running scripts / environments.
I would appreciate any useful links to articles or vidos.