Etcd should be upgraded to 3.5.24 to fix an etcd upgrade bug when going from 3.5 to 3.6

[recollir]

/kind bug

1. What kops version are you running? The command kops version, will display
this information.

1.34

2. What Kubernetes version are you running? kubectl version will print the
version if a cluster is running or provide the Kubernetes version specified as
a kops flag.

The bug is affecting an upgrade from Kubernetes version 1.33 to Kubernetes version 1.34

3. What cloud provider are you using?

AWS. But the bug should be independent of this.

4. What commands did you run? What is the simplest way to reproduce this issue?

In https://etcd.io/blog/2025/upgrade_from_3.5_to_3.6_issue_followup/ it is stated that upgrading etcd 3.5.X to 3.6.Y is potentially broken for etcd version <=3.5.23 and it is recommended to be on version 3.5.24 first.

kops 1.34 does the following right now

etcd 3.5.23 when building (or upgrading) a kubernetes cluster to Kubernetes < 1.34
etcd 3.6.5  when building (or upgrading) a kubernetes cluster to Kubernetes 1.34+

I guess we would want to bump etcd from 3.5.23 to 3.5.24 in kops 1.34 (and maybe also in the older still maintained versions from 3.5.21 to 3.5.24)

Additionally, a note in the changelog or upgrade documentation would probably be useful to make people aware that they should upgrade etcd first while staying on the same Kubernetes version before using kops 1.34 to upgrade kubernetes to 1.34/.

9. Anything else do we need to know?

I can attempt to "clone" PR #17639 to get 3.5.24 into the code base.