Add telemetry for offset lag

[LordSarevok]

What is the goal? how is it being implemented?

Open a PR to gather feedback about a feature that our team needs.

We are interested in the metrics from the point of view of the application about the status of the offsets of a topic and partition.

Our motivation is to have apart from visibility, alerts on this metric.

These metrics can be gathered from the broker, but there is no way to extend them with proper application context information, and for our use case it would be nice.

Please, I didn't write the test yet or the readme because I'm not sure if you will accept PRs like this (I saw no telemetry in the project, and maybe is there a reason for it) and also the approach itself, maybe due my lack of knowledge on the library I wrote the code in the wrong place.

Thanks!

[Argonus]

Hi, hmmm that looks interesting, but I'd like to see it as optional support. Not everyone is using telemetry, and most other packages I've seen is making such a think as optional configurable.

[LordSarevok]

Hi, hmmm that looks interesting, but I'd like to see it as optional support. Not everyone is using telemetry, and most other packages I've seen is making such a think as optional configurable.

Thanks for the fast reply! Do you know which of these packages can I use as an example? I can follow the same approach to avoid the dependency :D

Thank you very much! 🙇

[LordSarevok]

Maybe it was easier than I thought, Do you mean using optional: true on the dependency definition and relying on Code.ensure_compiled??

Again, thank you very much 🙇 !!

[Argonus]

Sorry, I've been off for two weeks. Hmmm, for example here we are using it for snappy vs snappyer. Also, sth simillary is in https://github.com/cabol/nebulex. Its perfectly fine for optional: true