The documentation for 1.10 release notes does mention this here
But if you just look at the relevant upgrade instructions, this change is not specifically called out, This could for example lead to a situation where a user has two containers in the pods where it crashloops as the newer containers cannot come up due to port conflicts:
$ k get pods -n cert-manager
NAME READY STATUS RESTARTS AGE
cert-manager-559f7bbfcd-9zrqg 1/2 CrashLoopBackOff 108 (32s ago) 8h
cert-manager-cainjector-6d46445566-65gh5 2/2 Running 0 8h
cert-manager-webhook-564c94659d-7fsmx 1/2 CrashLoopBackOff 108 (39s ago) 8h
We call out the OpenShift issue, so I think we should also call out this issue as it should be common enough due to helm's popularity.
The documentation for 1.10 release notes does mention this here
But if you just look at the relevant upgrade instructions, this change is not specifically called out, This could for example lead to a situation where a user has two containers in the pods where it crashloops as the newer containers cannot come up due to port conflicts:
We call out the OpenShift issue, so I think we should also call out this issue as it should be common enough due to helm's popularity.