On which service is it convenient to run a program with a highly floating load forever?

Question

Network administration

My program most of the time sluggishly monitors an external database with tasks. Makes a query to the database every few minutes. The program sleeps between requests - time.sleep(120) in python. At the same time, resource consumption (especially CPU) is very low.
Conventionally, once an hour, the program notices and reverses the task, while an external subprocess is launched, which heavily loads the CPU and requires more RAM.
The program runs under Linux (Ubuntu) 64bit and does not require a graphical interface.
To work, you need 10-100 gigabytes on the HDD for data storage. Not so much for active work, but for archiving.
Now she is spinning at my house. I want to run a script somewhere in the cloud and work with it over ssh.
Where can I place this program? Is there a system where I would pay for CPU and RAM as I consume them? It seems to me that I need some kind of AWS, but I can't figure out what exactly.

Answer 1

The question is how much are you willing to invest in learning the platform and adapting your application.
Possible options:
- the most fashionable, but with a serious learning curve - Kubernetes. I don't think it's worth it
- GCP App Engine, AWS Elastic Beanstalk, Heroku, etc. - a more realistic option
- just a cheap VPS (OVH, DigitalOcean, Linode, there are also cheaper ones) - the most primitive, but also the easiest

Answer 2

everything is like this: AWS or Azure
with a calculator, study the options - there is a bias either towards the CPU or HDD

Answer 3

I will support dimonchik2013 , but I will add that if you have such a small load and python, then maybe look towards serverless? Thus, it may turn out that it will even be shareware free.

Answer 4

in all, use fargate in my new service for launching containers