This is a tough one: It seems like we update span timestamps based on corrected timestamps for streamed fetch requests which are stored in a map. The logic itself is easy to port to client.on('processSpan') but the question is rather, if the map is populated as previously once the span ended. Given our span capturing and enqueuing logic is sync, and once enqueued we can't mutate spans anymore, this might be a timing issue we can't solve correctly.
This is a tough one: It seems like we update span timestamps based on corrected timestamps for streamed fetch requests which are stored in a map. The logic itself is easy to port to
client.on('processSpan')but the question is rather, if the map is populated as previously once the span ended. Given our span capturing and enqueuing logic is sync, and once enqueued we can't mutate spans anymore, this might be a timing issue we can't solve correctly.