easy

Results 56 comments of easy

IIUC: if the parent is sampled then the child is sampled.

If you want to avoid "inflationary" sampling of child spans, set the sampling probability to zero. Children of sampled parents will still get sampled.

Currently we sleep to let the stats aggregation and exporting happen, which isn't ideal.

I've been thinking about it but I don't have a good proposal yet. Related discussions: * https://github.com/census-instrumentation/opencensus-specs/issues/152 for flushing stats * https://github.com/census-instrumentation/opencensus-specs/issues/126 for flushing traces

I'm afraid not. I'm very limited in how much time I can spend on OpenCensus. Check out https://github.com/open-telemetry/opentelemetry-cpp/

I think these functions should return `grpc::Status` because almost all of the possible errors come from the RPC stack. The other kind of Status we have in OpenCensus is for...

This is one of the bullet points tracked in #244.

https://github.com/abseil/abseil-cpp/commit/bf29470384a101b307873b26d358433138c857fc - This somewhat large commit contains abseil's implementation of cmake install. We should take this into consideration when we do this for OpenCensus.

Enabling sampling after starting a span is mentioned in #218 also. A potential problem I can see is: if sampling_enabled is stored in the SpanContext, and the SpanContext has been...