Flaky test ContinuousProfilerTests.ExportAllocationSamples

[pellared]

When running nuke Workflow on macOS 15 with x64 I get very frequent failures from IntegrationTests.ContinuousProfilerTests.ExportAllocationSamples.

17:31:46 [DBG] [xUnit.net 00:16:38.54] IntegrationTests: --- FAIL: IntegrationTests.RabbitMqTests.SubmitsTraces(6.0.0) (68,17s)
17:31:46 [DBG] [xUnit.net 00:16:38.54] IntegrationTests: === RUN   IntegrationTests.RabbitMqTests.SubmitsTraces(7.0.0)
17:31:46 [DBG]   Niepowodzenie IntegrationTests.RabbitMqTests.SubmitsTraces(packageVersion: "6.0.0") [1 m 8 s]
17:31:46 [DBG]   Komunikat o błędzie:
17:31:46 [DBG]
17:31:46 [DBG] Missing expectations:
17:31:46 [DBG]   - "<no description> Instrumentation Scope Name: 'OpenTelemetry.AutoInstrumentation.RabbitMq', predicate is null: 'False'"

10:49:59 [DBG] [xUnit.net 00:21:27.83] IntegrationTests: === RUN   IntegrationTests.ContinuousProfilerTests.ExportAllocationSamples()
10:50:13 [DBG] [xUnit.net 00:21:41.21] IntegrationTests: --- FAIL: IntegrationTests.ContinuousProfilerTests.ExportAllocationSamples() (13,33s)
10:50:13 [DBG]   Niepowodzenie IntegrationTests.ContinuousProfilerTests.ExportAllocationSamples [13 s]
10:50:13 [DBG]   Komunikat o błędzie:
10:50:13 [DBG]    Resource attribute "os.build_id" was not found
10:50:13 [DBG] Resource attribute "os.description" was not found
10:50:13 [DBG] Resource attribute "os.name" was not found
10:50:13 [DBG] Resource attribute "os.version" was not found

Originally posted by @pellared in #4000 (comment)

It looks like that there is a race condition that the resource is configured after continous profiler is initialized.