I have a setup where there are multiple HAProxy instances involved within a request. I have set the
unique-id-header directive on the "outermost" HAProxy configuration, and it is now propagated through to the application over all the hops. However, the intermediate instances do not seem to understand that it could pick up the header, so the log format variable
%ID is not populated. If I set
unique-id-header again on those HAProxy instances, they add their own, new, header in addition to the existing one.
Is there some way to solve this? I could capture the header and have a different log-format on the other instances, but I'd prefer not to, if possible.