Fix duplicate instrument registration to return same instrument

[MrAlias]

Add a new cache type to handle generic lookup functionality. Wrap this new type in an new instrumentCache to be used by the resolver when resolving Aggregators. If an instrument has already been resolved, it returns those aggregations or an error.

The new cache type is defined generically so its can be unified with the meterRegistry, which duplicates the functionality. That unification is left to a follow-on PR.

---

• Fix Duplicate instrument registration generates an error instead of a valid instrument #3229
• Supersedes Adds a cache to the metric insrument pipeline. #3181

[codecov]

Codecov Report

Merging #3238 (1d4f1d2) into main (aca054b) will increase coverage by 0.1%.
The diff coverage is 100.0%.

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #3238     +/-   ##
=======================================
+ Coverage   77.3%   77.4%   +0.1%     
=======================================
  Files        159     160      +1     
  Lines      11184   11228     +44     
=======================================
+ Hits        8651    8699     +48     
+ Misses      2335    2331      -4     
  Partials     198     198             

Impacted Files	Coverage Δ
sdk/metric/cache.go	100.0% <100.0%> (ø)
sdk/metric/meter.go	100.0% <100.0%> (ø)
sdk/metric/pipeline.go	96.4% <100.0%> (+0.4%)	⬆️
sdk/trace/batch_span_processor.go	81.9% <0.0%> (+1.7%)	⬆️

[MrAlias]

This does not resolve #3240. I'm looking into how feasible it would be to resolve that as well here, but its resolution might need to be a follow-on PR.

[MrAlias]

This does not resolve #3240. I'm looking into how feasible it would be to resolve that as well here, but its resolution might need to be a follow-on PR.

It does look like we can cache lower in the computation processing to handle #3240: MrAlias#661. I'm going to close this so it is not reviewed, and refactor a more complete solution for another PR.