[Tracking] - Code Health Overhaul

[rauljordan]

💎 Issue

Background

Over the past few months, our team has been a lot more focused on code health. Namely, we want to adopt successful software engineering practices into our organization to help Prysm succeed as a project, and make our life easier as developers. Better code health means more testable, concise, and simpler code that is easier for contributors to add value to and also for fewer bugs to arise. We have recently published Software Design Principles in Go, and now have a direction for how to start implementing these ideas in Prysm. This is a tracking issue for all code health related tasks grouped by domain.

Description

Separation of Concerns

• Avoid leaking all instances of CLI flags, cli.Context beyond the main packages in Prysm
• Use function options for configuring services in Prysm

Encapsulation

• Encapsulate all feature flag conditionals as close as possible to what they toggle
• Encapsulate nil checks across the repository and return package-level errors such as ErrNotFound Encapsulate Nil Checks #9454

Single Responsibility Principle

• Use functional validation pipelines for reusable validity checks on data structures in Prysm

Open/Closed Principle

• Currently, we use type assertions on the various kinds of Keymanager implementations across Prysm, which defeats the purpose of using interfaces. This is a violation of the open/closed principle. Adding a new keymanager implementation to Prysm shouldn't mean we have to modify the rest of the codebase to add new type assertions for it

[rauljordan]

Related: #9883

[rkapka]

Related: #7492

[terencechain]

Let's close this and open this in a separate issue