Chances of refactoring shadow?

[wi1dcard]

As the maintainer of shadow instances, I found that we always have to somehow carefully "craft" the shadow configuration and runtime environment to ensure that shadow can run without any strange issues.

From what I can tell now, there are at least 4 different types of problems that need to be resolved:

• Missing docs and confusing, bizarre behavior: Question and Feature Request: -v or RUST_LOG, which one is prioritized? #142
• Doesn't consider the deployed environment on production: Feature Request: Do not use paths under $HOME as the default storage path #128, API server should not start threads based on the number of cores #115, Feature Request: Disable all API request logs optionally #129, Generating DAG consumes all the CPU quota #158
• Bugs: Cannot import #157, Potential memory leak #160
• Bad tech stack selection (and potentially a bug): Consider the impact if have to re-import the data from geth #159

Lack of documentation, little consideration of code robustness, bad architecture design generate too heavy load to the developers and maintainers. This is also a very important reason why we didn't have a staging environment for shadow. I believe it can be even more critical once if we decide to allow others to run a shadow instance or make the bridge public.