feat: Switch prints to tracing

[itegulov]

Fixes #26

I chose tracing since this is what nearcore is using and due to its extensive features such as layered spans and subscription mechanisms that can be useful for monitoring further down the line. As per conversation under the issue, it does not necessarily tie us to tokio ecosystem either.

Unfortunately, the sandbox logs are being produced by a dynamically downloaded binary file, so they do not respect the logging configuration (RUST_LOG, RUST_LOG_STYLE and others). See #53. This means that for now users will see logs being printed in an inconsistent manner like this:

2022-01-12T01:05:05.814561Z  INFO workspaces: Starting up sandbox at localhost:15562
2022-01-12T01:05:05.814601Z  INFO workspaces: Starting up sandbox at localhost:15965
Jan 12 12:05:05.817  INFO neard: Version: trunk, Build: 2c9375ee, Latest Protocol: 48
Jan 12 12:05:05.817  INFO neard: Version: trunk, Build: 2c9375ee, Latest Protocol: 48
Jan 12 12:05:05.817  INFO near: Generated node key, validator key, genesis file in /tmp/sandbox-15965
Jan 12 12:05:05.817  INFO near: Generated node key, validator key, genesis file in /tmp/sandbox-15562
2022-01-12T01:05:05.817696Z  INFO workspaces: Started sandbox: pid=12175
2022-01-12T01:05:05.817866Z  INFO workspaces: Started sandbox: pid=12176

[itegulov]

Also, I have not touched any of the examples, but it might be good idea to refactor at least one (or all?) of them to show that workspaces can be used with either log or tracing. @ChaoticTempest WDYT?

[ChaoticTempest]

it might be good idea to refactor at least one (or all?) of them to show that workspaces can be used with either

@itegulov Definitely a good idea, wanna try refactoring the spooning example? There's only a couple of println calls in there

[itegulov]

@ChaoticTempest Done!