Istio 1.4.0 tracing service moved from port 80 to 9411

[Stono]

Bug description
In istio 1.4.0, tracing.istio-system has moved from port 80 to 9411. There is already a zipkin.istio-system service in the namespace on 9411. Both services point to the same workload (Jaeger).

This is a breaking change for us so just wondering why it was made and if it was intentional, it should have been in the upgrade notes (as we have upstream systems talking to tracing.istio-system).

Expected behavior
Consistent service ports between releases or details in release notes explaining the changes.

OR if you're really honouring semver releases, 1.4.0 shouldn't have breaking changes from 1.3.x and a different service port could probably be classed as breaking.

Steps to reproduce the bug

Version (include the output of istioctl version --remote and kubectl version and helm version if you used Helm)
1.4.0

How was Istio installed?

Environment where bug was observed (cloud vendor, OS, etc)

[Stono]

Associated commit appears to be 9e0887f#diff-264788198be8f0b9371c31fecce6d199 and issue #17204

[sdake]

Nodeport more trouble then it is worth... These two PRs need a revert. The author is free to resubmit the work.

[sdake]

Had a deeper look at the PR. With istioctl dashboard jeager for example, none of these config options are needed across any of the third party component charts. We should consider their removal.

@Stono for the moment, I am going to make a small surgical change that should exhibit the original behavior.

[Stono]

Thanks @sdake - no huge urgency on my part, I've patched around it locally. Just don't want other users getting caught out when they upgrade to 1.4.0.