Joshua MacDonald

I came to this issue, realizing a related problem exists. When the `exporterhelper` code has queuing enabled and encounters a full queue, it will drop the data and log a...

I was confused, in the comment above, by the issue identified here: https://github.com/open-telemetry/opentelemetry-collector/pull/8674#pullrequestreview-1672721521. Exporters that use exporterhelper are counting `enqueue_failed` metrics.

With respect to the question about taking a dependency on the go-expohisto reference implementation: It's well documented that I briefly succeeded at placing this code into the otel-go metrics SDK....

@jack-berg PTAL

Currently, the logging exporter prints exponential histograms incorrectly according to the specification. I would like it fixed.

Discussed in the 11/17/22 Sampling SIG. This would be addressed by the proposal dicsussed in https://github.com/open-telemetry/opentelemetry-specification/issues/1413, which would take advantage of 7 definite bytes of randomness per TraceID.

I disagree with the effort to protect users from being foolish. I support the idea of removing zero-valued enums, but not removing non-zero values. The FLAG_NO_RECORDED_VALUE has disappeared, which I...

More in https://github.com/open-telemetry/opentelemetry-specification/issues/2746#issuecomment-1319259109

@PeterF778 would you agree that would be possible for an SDK to do what you described by fixing the 7 bytes of random TraceID and then generating multiple correlated TraceIDs...

I support the call to stabilize the exponential histogram as we have it. I agree there are probably useful additions that we can find, but we should be able to...