John Howard

> I see, I suppose I was still looking at it from a GitOps perspective. If one were to use not-GitOps (HumanOps?) that would totally be production-usable. https://blog.howardjohn.info/posts/operators/index.html#gitops for my...

Its fine to say this won't be fixed - I have no interest in criticizing the projects decisions - but I did want to clarify: the issue was not "Helm...

Given most of this has landed and OC is deprecated I don't think there is anything left to do. OTEL is much better here: https://blog.howardjohn.info/posts/zero-alloc-otel/

Another option: requiredDuringSchedulingIgnoredDuringExecution matching us to co-schedule with k8s-api=istio-cni. The problem, I think, is that doesn't check Readiness of CNI

problem is you can only have 1 scheduler so if they alredy defined one we will break them On Thu, Dec 8, 2022 at 11:14 AM Zhengzhe Yang ***@***.***> wrote:...

https://github.com/kubernetes/kubernetes/pull/113274 is like scheduler idea kind of. But it's alpha I think

I am not sure we have an issue for waypoints, or its at least no worse than sidecars? At least from my initial thinking Case 1. we are spinning up...

New ideas: fixing it in the primary CNI (e.g. cilium) itself. Option 1: make the CNI config have an empty plugin spec (like `{type:istio-cni}`). Pods will be blocked with an...

> IIRC Cilium already supports supplying a template /etc/cni/net.d conflist during install. So (Op. 1) can already probably work. I think this is just a bool, so we would need...

> Every user that provisions their own nodes can do that today with Istio, or any number of other things, correct? Either you have the capability to provision nodes to...