Joshua MacDonald

Updated for v0.62.0.

I believe all but one of these are addressed in #36.

On further investigation, there are two categories of test failure that are still happening here after #36. First, there are tests that use `time.Sleep()` that may break under heavy load,...

Sure! I ran overnight testing and was able to reproduce a few test failures. Every failure I looked into, however, just appeared to be a _minor_ issue in the test,...

Here's the most frequent one: ``` --- FAIL: TestServerResendClientInflightBackoff (0.00s) server_test.go:2208: Error Trace: server_test.go:2208 Error: Not equal: expected: []byte{0x3a, 0xe, 0x0, 0x5, 0x61, 0x2f, 0x62, 0x2f, 0x63, 0x0, 0xb, 0x68,...

The OTLP exporter is welcome to skip exporting metrics with no data points, I feel. Does this need to be specified? I also feel that Delta temporality is the correct...

The requested OTLP clarification is in https://github.com/open-telemetry/opentelemetry-specification/issues/2715. Given that we support duplicate instrument registration, it's not clear that calling "create" will actually create anything. However, I think the use of...

@legendecas There are at least four issues being discussed here and in #3105. If you're using delta temporality, you should be able to avoid memory buildup regardless of which attributes...

Thanks @legendecas. You are correct. My statement about delta temporality only applies to synchronous instruments. This is the sense in which Lightstep would like to support a temporality preference named...

To get Lightstep onto using the OTel APIs I had to add several off-spec metrics features features (mainly the synchronous gauge aggregation), which would ideally be expressed via hints. The...