Add benchmarks for the prometheus bridge

[dashpole]

I wanted to get some rough numbers for how much overhead the prometheus bridge + OTLP exporter adds to pure Prometheus instrumentation, and also compare that with the Prometheus receiver vs OTLP receiver in the collector.

They aren't the best benchmarks (I used things like top and docker stats to measure), but the results broadly show what I expected, which is that is is much less expensive to do the translation client-side with the bridge, than server-side with the Prometheus Receiver.

My sample code: prometheus/client_golang@main...dashpole:client_golang:benchmarking

cc @gouthamve @logicalhan

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 61.9%. Comparing base (b343370) to head (daecb73).

❗ Current head daecb73 differs from pull request most recent head 9b9d10a. Consider uploading reports for the commit 9b9d10a to get more accurate results

Additional details and impacted files

@@          Coverage Diff          @@
##            main   #5364   +/-   ##
=====================================
  Coverage   61.8%   61.9%           
=====================================
  Files        186     186           
  Lines      11388   11388           
=====================================
+ Hits        7049    7051    +2     
+ Misses      4137    4136    -1     
+ Partials     202     201    -1     

see 1 file with indirect coverage changes