[prometheus] Stabilize Prometheus -> OTLP: Resource Attributes

[dashpole]

What are you trying to achieve?

Stabilize https://opentelemetry.io/docs/specs/otel/compatibility/prometheus_and_openmetrics/#resource-attributes

Open questions:

1. Should we keep translating job and instance to service.name and service.instance.id by default?
2. Should we keep dropping target_info and converting it to resource attributes?

I think the answer to (2) is yes. Even if we were to no longer export target_info from prometheus exporters, it seems like the receiver portion should still support turning those into resource attributes.

We should discuss (1). There have been cases of users being confused as to why they can't use processors on the job and instance labels. But overall, i'm not convinced that breaking users would be worth it at this point regardless of our opinions on which was less confusing.