feat(libs): add libs

[ShaunSHamilton]

1^st in stack

#1
├── #2
├── #3
├── #4
├── #5
├── #6

[ojeytonwilliams]

Don't feel like you need to explain in detail (pointing me to the docs would be great), but is the shutdown analogous to fastify's? i.e. when you start shutting down new then connections are rejected with a bad status (503, say) and (depending on config) existing connections get killed at some point?

[ShaunSHamilton]

The main thing happening here is:

fn main() -> Exit {
    // ...stuff that does not matter
    // `_sentry` guard created
    let _sentry = runtime::telemetry::init(runtime::release!(), &config.telemetry);

    runtime::serve(config.listen_addr, app::app())
    // `_sentry` guard dropped
}

Once _sentry "goes out of scope", the Drop implementation is invoked. If we do not explicitly set Sentry's timeouts, the defaults are used, and this guard is held on for annoyingly long - Docker will force kill, before graceful shutdown.

Refs:

• https://doc.rust-lang.org/rust-by-example/scope/raii.html#destructor
  • Short explanation of Drop as RAII
• https://github.com/getsentry/sentry-rust/blob/4b66d3c2c123bd6c6f1be694fce39e1d9c9d8d32/sentry/src/init.rs#L31-L44
  • Actual sentry Rust code called

---

As for your actual question, I am not sure this can be compared to Fastify, because this code is part of the CLIENT.

[ojeytonwilliams]

Okay, thanks. I'm sure this will become clear, but I saw this being used in the server file and so my mind didn't leap to client!