observability: metrics, tracing

[srenatus]

When run as a server, it would be nice to feature generic (go runtime data) and specific metrics (source update, bundle push).

Also, since it involves multiple other services, tracing would be a neat feature. Basically OpenTelemetry, and some instrumentation, could go a long way.

Ideally, these would be stashed in extra modules to keep the main (go) module slim.

[srenatus]

🎉 With #118, we've got a first set of metrics to explore. Let's figure out collectively if this is what we need for operating OCP, or where further insights are needed. 🔍 📝