Timing issues with different systems involved

[DanielMSchmidt]

Hi there,

First of all big thanks for maintaining this project! We are adding zipkin to our application and experience a problem with the time sync between our frontend running in a browser and our backend systems running on AWS or in a kubernetes cluster. We got spans that look like this:

I ensured that the span is only finished when the request comes back from the server and the times indicate that it should be fine, except for a some second offset between both systems. Is there some way I can "sync" both zipkin clients with each other to rely on the same timing? Is this not a common issue?

[DanielMSchmidt]

/cc @levito & @garthenweb

[codefromthecrypt]

looks related to this? openzipkin/zipkin#1492 (comment)

[mpetazzoni]

looks related to this? openzipkin/zipkin#1492 (comment)

I think you meant openzipkin/zipkin#1492 (comment) (my comment just above it), but yes it looks very similar. The hosts participating in the trace have slightly offset clocks. I'm not sure the dependency-linking discussed in openzipkin/zipkin#1480 would help there, unless I'm misunderstanding something.

[codefromthecrypt]

This issue was moved to openzipkin/zipkin#1839

[codefromthecrypt]

moved this one as nothing special would happen in this repo