[GitHub] [skywalking] wu-sheng commented on issue #2653: Improve Zipkin<->SkyWalking integration

[GitBox <gi...@apache.org>]

wu-sheng commented on issue #2653: Improve Zipkin<->SkyWalking integration
URL: https://github.com/apache/skywalking/issues/2653#issuecomment-538449350
 
 
   > Let me confirm, the only reason that we need to match the spans in zipkin but no need to match the spans in skywalking native spans is because, skywalking native spans' spanid can tell the sequence of spans made from the single segment, while zipkin can not tell the SERVER spans' sequence without knowing the timestamp of CLIENT spans. Is that correct?
   
   This is for one, the other is, you can't tell the `server` span's parent is `client` span, because you could have `server` span as `server` span's parent, such as `Tomcat` span as `SpringMVC` span's parent.
   
   > I have a quick way solving the problem. I can setup a proxy layer for the span receiver, and based on the trace id, we can route the span to the target node for processing. The theory is kind of consistent hash ring. Do you agree for the quick solution?
   
   You could, but how much memory do you expect to cost, and how long timeout should be set, because you never know, there will or will not more span coming.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services