Improve documentation

[idantene]

See slack messages starting here but also elsewhere by searching for docs or documentation.

ClearML is great, but I often find myself rummaging through the GitHub code for more information, or asking silly questions on Slack simply because the documentation is somewhat lacking.
To be precise, the documentation is:

1. Rather complete in terms of SDK documentation
2. Very lacking in terms of usage documentation
3. SDK documentation is very messy

I think ClearML and the users would benefit greatly from mirroring the numpy ecoysystem's documentation style.
That is, documentation should:

1. Include an overview of how different elements/SDK instructions interact with each other (agent, queue, remote execution, configuration files, artifacts, storage manager, etc)
2. SDK documentation should be easy to read, easy to find formatting issues, easy to search through (e.g. one long page for Task is simply difficult to navigate through).

As an example, consider pandas documentation:

1. A separate structured user guide with common tips, usability, best practices - https://pandas.pydata.org/pandas-docs/stable/user_guide/index.html
2. SDK documentation, where each function is its own page, e.g. https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.read_csv.html
3. Examples for SDK (and user guide) live both in GH but also in "code" in their respective pages, with multiple snippets explaining different use-cases

[ainoam]

Thanks for the detailed feedback @idantene!

ClearML's documentation is constantly being added to and improved and your feedback goes a long way in pointing us in an effective direction.
As a side-note, appreciate if you can make the issue title a little more precise, as "Improve Documentation" will probably always be applicable in one manner of another :)