Prometheus Metrics Documentation

Prometheus metrics are documented only in the metrics endpoint in about half a sentence. I couldn't find more info in the docs.
These metrics weren't clear for me:
**turn_traffic_peer_***
What are these and how are they different from turn_traffic_* counterparts? From what I could see, turn_traffic_peer_*s were a little less than turn_traffic_*s.
**stun_binding_response**
Description says "Outgoing STUN Binding responses". Does this mean number of successful responses? I couldn't see this metric being reported even though I made a couple of STUN requests (stun_binding_request metric was reporting 80) and I could see the successful response.
**process_start_time_seconds**
Description says "Start time of the process since unix epoch in seconds", But when I run it, reports something like 2731941 which is obviously wrong (2.7 million seconds is about two years). I checked the date on the container and it was okay.

Start with the Prometheus metrics endpoint and compare the definitions and observed values for turn_traffic_peer_*, turn_traffic_*, stun_binding_response, and process_start_time_seconds. Check the existing metrics documentation and clarify each metric's meaning, counting conditions, and reported time value; done means the documented behavior matches the endpoint and observed results.

Written by the indexing model from the issue text.