Feature request: add Sentry support for reporting errors

[mirceaulinic]

I ❤️ Sentry and I'm using it in almost every project / app that runs in background that can silently fail, which many times turned out to be serious issues - that would be unnoticed otherwise.

The integration is relatively simple (see https://docs.sentry.io/platforms/rust/). I'd be happy to add it myself, however I'm yet to learn Rust. Anyway, I'm opening this for now mainly for hearing your thoughts.

Usually the Sentry DSN is read from an environment variable (e.g., SENTRY_DSN) and / or CLI arg (e.g., --sentry-dsn).

[partim]

I’ve read up a bit on Sentry – and it wasn’t quite what I expected. So, as a bit of clarification. Do I understand correctly that essentially Routinator would submit panics as events into Sentry and offer it as a logging backend?

[mirceaulinic]

Correct @partim - with the distinction that I wouldn't call it logging but rather monitoring; and, yes, monitoring errors / panics as events that provide a detailed trace from where the error started which helps identifying the root cause easily.

[partim]

Sounds very interesting, indeed! I’ll do a bit more research into the best way to include support.