I was exploring ease of use and more automated solution and tried to do horizontal scaling and vertical scaling as a POC.
Horizontal Scaling:
I have implemented horizontal scaling using SCALESET(azure service used)
Problem: it’s not working as expected, although I have set up logic like whenever CPU hits 60%, SCALSET has to spin up 2 more VM machines but it didn’t happen.
Vertical Scaling:
I am in process of vertical scaling on the VM based on CPU usage.
Problem Statement:
To convert the files from .dat to .parquet using COM DLL with python code(batch processing , not a web traffic)
Questions:
- Which one is the correct fit for my problem- vertical/horizontal scaling?
- Does horizontal scaling support in this scenario for internal load and how do we need to configure, I tried and it dint work out for some reason.
- If we choose vertical scaling, does the COM DLL works with high processing or not?