Prometheus exporter no longer collects metrics after shutdown

[dashpole]

Fixes #4032.

Before the fix, the test case fails with:

--- FAIL: TestShutdownExporter (0.00s)
    exporter_test.go:859:
        	Error Trace:	/usr/local/google/home/dashpole/go/src/go.opentelemetry.io/opentelemetry-go/exporters/prometheus/exporter_test.go:859
        	Error:      	Received unexpected error:
        	            	collected metric "target_info" { label:{name:"service_name" value:"unknown_service:prometheus.test"} label:{name:"telemetry_sdk_language" value:"go"} label:{name:"telemetry_sdk_name" value:"opentelemetry"} label:{name:"telemetry_sdk_version" value:"1.19.0"} gauge:{value:1}} was collected before with the same name and label values
        	Test:       	TestShutdownExporter
FAIL
FAIL	go.opentelemetry.io/otel/exporters/prometheus	0.055s

[codecov]

Codecov Report

Merging #4648 (52de753) into main (d18277e) will increase coverage by 0.0%.
The diff coverage is 75.0%.

@@          Coverage Diff          @@
##            main   #4648   +/-   ##
=====================================
  Coverage   81.3%   81.4%           
=====================================
  Files        225     225           
  Lines      17725   17728    +3     
=====================================
+ Hits       14427   14431    +4     
  Misses      2998    2998           
+ Partials     300     299    -1     

Files	Coverage Δ
exporters/prometheus/exporter.go	84.0% <75.0%> (+0.4%)	⬆️