Reduce file system usage

[olegbespalov]

What?

Currently, the analysis uses file system widely, creates artifacts and cleans it. It's okay, but in theory we can do some of these if not the all operations in memory, reducing the footprint for the system where we execute.

One of the examples of that movement is #69 pull request, but in theory there could be more

I'm opening issue here, but in theory it also affects the k6pack, etc

Why?

It reduces the footprint for the system where we execute, also makes the analysis more robust for issues, like if something went wrong (panic), there is no need to organize the careful clean-up