Add test to validate log shutdown

[cijothomas]

Related to #1514, not quite the finish, but just a initial test.

[codecov]

Codecov Report

Attention: Patch coverage is 90.00000% with 8 lines in your changes are missing coverage. Please review.

Project coverage is 68.3%. Comparing base (a8b308a) to head (26e8e64).
Report is 2 commits behind head on main.

Files	Patch %	Lines
opentelemetry-sdk/src/logs/log_emitter.rs	90.0%	8 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #1620     +/-   ##
=======================================
+ Coverage   67.8%   68.3%   +0.4%     
=======================================
  Files        139     139             
  Lines      19687   19806    +119     
=======================================
+ Hits       13363   13530    +167     
+ Misses      6324    6276     -48     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[lalitb]

Thanks. Nice way to validate the scenario described in #1514 :)

[lalitb]

In a diff. test, I intend to test that it is indeed causing the effect of flush(), but not necessarily by calling flush() itself.

Looking into the batch processor code, it seems to be doing the flush before shutdown. But good to validate this in separate test.

opentelemetry-rust/opentelemetry-sdk/src/logs/log_processor.rs

Lines 212 to 220 in 5ef3fd3

	BatchMessage::Shutdown(ch) => {
	let result = export_with_timeout(
	config.max_export_timeout,
	exporter.as_mut(),
	&timeout_runtime,
	logs.split_off(0),
	)
	.await;

[cijothomas]

@hdost @lalitb Could you re-review? I added a separate-thread-logger scenario too.

[cijothomas]

In a diff. test, I intend to test that it is indeed causing the effect of flush(), but not necessarily by calling flush() itself.

Looking into the batch processor code, it seems to be doing the flush before shutdown. But good to validate this in separate test.

opentelemetry-rust/opentelemetry-sdk/src/logs/log_processor.rs

Lines 212 to 220 in 5ef3fd3

	BatchMessage::Shutdown(ch) => {
	let result = export_with_timeout(
	config.max_export_timeout,
	exporter.as_mut(),
	&timeout_runtime,
	logs.split_off(0),
	)
	.await;

yes. that'd be a targetted test for processors to ensure shutdown() gives same effect as force_flush().

[lalitb]

Nice :)

[cijothomas]

@hdost @TommyCpp I'll go ahead and merge this to continue working on this. Please let me know if you have any feedback on this approach.