Fix flaky OTLP exporter reconnect test

[MrAlias]

The tests that check the OTLP exporter will reconnect wait for the reconnect loop to complete, in theory. However, they do not yield the active goroutine scheduling. The reconnect loop is in its own goroutine meaning it is unlikely for that loop to be re-entered, especially on slow systems. This updates the tests call runtime.Gosched when waiting for the reconnect loop and yield the scheduling to other goroutines.

This does not clean up the duplicate testing (TestNewExporter_collectorConnectionDiesThenReconnectsWhenInRestMode seems to be a paired down version of TestNewExporter_collectorConnectionDiesThenReconnects), which I'm planning to address in a follow up PR. Because of this, the reuse of the "wait" function is just duplicated instead of abstracted into its own declaration.

Related to #1527

[codecov]

Codecov Report

Merging #1814 (7cb51a1) into main (b09df84) will decrease coverage by 0.0%.
The diff coverage is n/a.

@@           Coverage Diff           @@
##            main   #1814     +/-   ##
=======================================
- Coverage   78.6%   78.6%   -0.1%     
=======================================
  Files        134     134             
  Lines       7158    7158             
=======================================
- Hits        5632    5630      -2     
- Misses      1276    1278      +2     
  Partials     250     250             

Impacted Files	Coverage Δ
exporters/otlp/otlpgrpc/connection.go	88.7% <0.0%> (-1.9%)	⬇️

[MrAlias]

This looks like it has not completely resolved this issue: https://github.com/open-telemetry/opentelemetry-go/pull/1814/checks?check_run_id=2346462483

I think it is a step forward and should be merged still, but will remove the linked issue as being resolved.