refactor(cache): optimize concurrency with RWMutex for cache operations

[flc1125]

Although it is an optimization, whether this optimization is worth doing requires consideration.

package main

import "go.opentelemetry.io/otel"

var meter = otel.Meter("xxx")

func main() {
	meter.Float64Counter("xxxx")
}

In most cases, examples like this show that otel.Meter is typically defined once and not called multiple times. This implies that the cache hit rate itself is extremely low. Consequently, the performance overhead between RWMutex and Mutex calls is almost negligible.

If this scenario holds true, the current PR can be ignored.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Please upload report for BASE (main@be4b9e7). Learn more about missing BASE report.

Additional details and impacted files

@@          Coverage Diff           @@
##             main   #6911   +/-   ##
======================================
  Coverage        ?   82.9%           
======================================
  Files           ?     261           
  Lines           ?   24264           
  Branches        ?       0           
======================================
  Hits            ?   20129           
  Misses          ?    3758           
  Partials        ?     377           

Files with missing lines	Coverage Δ
sdk/metric/cache.go	100.0% <100.0%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[MrAlias]

Please include comprehensive evidence in the form of benchmarks and/or profile data to support this change.

This has been evaluated in development and determined to be less performant in the common use-cases. Direct evidence is required for these changes to be considered.

[pellared]

Isn't RWMutex having a bigger memory footprint?

• https://github.com/golang/go/blob/d05825468921f8d0e54c81f59330d3b263057a54/src/internal/sync/mutex.go#L17-L23
• https://github.com/golang/go/blob/d05825468921f8d0e54c81f59330d3b263057a54/src/sync/rwmutex.go#L39-L45

Are there any benchmarks that proves that it is an optimization?