Marco Voelz

I know I'm kind of late to the party and discussion, but have y'all considered switching on [`autopath`](https://coredns.io/plugins/autopath/) for CoreDNS? It does require [`pods verified`](https://coredns.io/plugins/kubernetes/), (which might not be what...

> I would generally recommend to do all pod-specific/runtime-related mutations on pods themselves instead of mutating higher-level controllers This is also my opinion, especially when there is already an existing...

Working with `ctr` continues to be a pain, every time we're looking at `containerd` issues I see people being tempted to bring in other tools (`nerdctl`, for example) so while...

Do we need to tackle this now instead of doing #4673 first, such that we can simply remove that code? Is there a benefit in touching this, knowing that it...

> Thinking again on this topic, Gardener can configure both [kubelet](https://github.com/gardener/gardener/blob/744e0ff5db4f7f58d70a2f43546d379303f8a2cf/pkg/operation/botanist/component/extensions/operatingsystemconfig/original/components/kubelet/config.go#L54) and containerd [(similarly to how it control the pause image)](https://github.com/gardener/gardener/blob/744e0ff5db4f7f58d70a2f43546d379303f8a2cf/pkg/operation/botanist/component/extensions/operatingsystemconfig/original/components/containerd/templates/scripts/init.tpl.sh#L9-L12) to use the wanted cgroup driver, so from this...

@vpnachev do you have some thoughts on the above discussion? It may be that we have some time to pick this up in the near future.

I understand that this is only a prerequisite for enabling `cgroupsv2`, but please be aware that additional things need to be done before switching the cgroup: - ensure that we're...

@elmiko > what would happen if a user downgrades their kubernetes version with this option in use on their VPAs I'm not sure how this is related to kubernetes versions....

Thanks for your thoughts on this, @jbartosik! My understanding is that you're talking about two different alternatives to my proposal here, so let me try to address them one-by-one. >...

> The API is a little bigger (adding a new message instead of a new enum value) but I hope that implementation should take about the same amount of work....