You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In this line, an X-Correlation-ID is always added based on the SpanID. Not sure what we need this for, but it can cause problems in scenarios where an X-Correlation-ID is already set beforehand. For instance, tracing tools may get confused as to which ID to follow. That said, going from req.Header.Add() to req.Header.Set() won't fix the problem, as the existing ID would be overwritten and tracing would be corrupted.
IMHO, there should be only one X-Correlation-ID, so at least it should only be set if there isn't one already. Probably even better to not use X-Correlation-ID here and go with another header instead.
The text was updated successfully, but these errors were encountered:
The X-Correlation-ID is always added to outbound requests' headers. It propagates the trace ID.
What do you mean by tracing tool? Because this header is not set on requests that are sent to tracing tools like Jaeger or Zipkin.
I don't know what was the idea behind adding it since the opencensus library should already take care of this.
@bastianccm can provide us maybe with more information :)
In this line, an X-Correlation-ID is always added based on the SpanID. Not sure what we need this for, but it can cause problems in scenarios where an X-Correlation-ID is already set beforehand. For instance, tracing tools may get confused as to which ID to follow. That said, going from
req.Header.Add()
toreq.Header.Set()
won't fix the problem, as the existing ID would be overwritten and tracing would be corrupted.IMHO, there should be only one X-Correlation-ID, so at least it should only be set if there isn't one already. Probably even better to not use X-Correlation-ID here and go with another header instead.
The text was updated successfully, but these errors were encountered: