logging: Improve logging and async debugging experience

[justinrubek]

We are currently using the tracing crate for logging, but there is potential for adding on to this for development. Especially as the application gets more complex and uses increasing amounts of async code this is becoming more difficult. When working on #63 it felt like there were major slowdowns at times because of this. I decided to learn more about the tracing library and immediately they offer a few suggestions:

• Tokio console
  • basically an async debugger
  • included in nixpkgs
• https://opentelemetry.io/

I do not have experience with any of these, but they look like they could be very useful and we could probably pick them up fairly simply. Integrating these tools, learning how to use them, writing a quick-guide, and making a pass on tracing could help significantly

[justinrubek]

Linked branch tokio-console. It's a fairly simple change but it seems that the console is limited. I can't find a way to look at the logs or see much of anything that would be used for basic debugging. There is some information to be had there so it's not entirely useless but it certainly is not what I'd expected.

It looks like the opentelemetry route may have promise. This needs some more setup and will have to be explored further. Since enabling tokio-console required configuring tokio_unstable I do not intend to open it in a PR at this time.

In the meantime, if you want to try out the console everything should be available in the nix devshell. The console subscriber is configured in the logutil crate so it is already wired into the server: cargo run --bin glaredb -- -v server and tokio-console