The v1.45.0 ThreadPoolExecutor shutdown wait=False change in opentelemetry.sdk.resources.get_aggregated_resources is triggering failures in leaked thread detector in unit tests

[vkruglik-aka]

Reference to change:

opentelemetry-python/opentelemetry-sdk/src/opentelemetry/sdk/resources/__init__.py

Line 776 in 4f0fcfa

executor.shutdown(wait=False, cancel_futures=True)

The v1.45.0 ThreadPoolExecutor shutdown wait=False change in opentelemetry.sdk.resources.get_aggregated_resources is triggering failures in leaked thread detector in app unit tests.

The regression introduced in v1.45.0:

1. Resource.create() call in MeterProvider.init() that spins up a background ThreadPoolExecutor to run resource detectors in parallel.
2. The 1.44.0 version of the sdk was using the ThreadPoolExecutor in a context manager expression, which joins all of its threads on exit from context.
3. However, version 1.45.0 doesn't use the context manager and instead shuts down the thread pool explicitly with wait=False, which bypasses the thread join calls resulting in the race condition with the thread leak detection in our test suite.

[akshat-lakhera]

Hey @nickstenning , I would like to work on this, can you assign me this issue.