Lambda

It's actually still tail based sampling, but it's happening on near, I guess, export time in process in the Lambda. So we don't have to retroactively go scramble to turn on a particular sampling ratio and try and get data. Different from the tail sampling that you'd either send the child spans or you wouldn't. If the child spans aren't being sent, you're getting all those fields rolled up to the root span. The root spans that you're keeping alone have a lot more data then if the child spans were created and then discarded later.