otlptracegrpc otlpmetricgrpc: Handling RESOURCE_EXHAUSTED

[pellared]

Description

From specification:

The client SHOULD interpret RESOURCE_EXHAUSTED code as retryable only if the server signals that the recovery from resource exhaustion is possible.
This is signaled by the server by returning a status containing RetryInfo.
In this case the behavior of the server and the client is exactly as described in OTLP/gRPC Throttling section. If no such status is returned, then the RESOURCE_EXHAUSTED code SHOULD be treated as non-retryable.

Current behavior

opentelemetry-go/exporters/otlp/otlpmetric/otlpmetricgrpc/client.go

Lines 175 to 192 in fc0a461

	// retryable returns if err identifies a request that can be retried and a
	// duration to wait for if an explicit throttle time is included in err.
	func retryable(err error) (bool, time.Duration) {
	s := status.Convert(err)
	switch s.Code() {
	case codes.Canceled,
	codes.DeadlineExceeded,
	codes.ResourceExhausted,
	codes.Aborted,
	codes.OutOfRange,
	codes.Unavailable,
	codes.DataLoss:
	return true, throttleDelay(s)
	}
	// Not a retry-able error.
	return false, 0
	}

RESOURCE_EXHAUSTED is always treated as retryable.

Expected behavior

Follow the specification.