Julian Tölle

That would be the `CloudNodeLifecycleController` in `kubernetes/cloud-provider`: https://github.com/kubernetes/cloud-provider/blob/8fe710fc4036e9992a88a5c89fdf9eaf4987b56f/controllers/nodelifecycle/node_lifecycle_controller.go#L153-L179 This in turn calls `InstanceV2.InstanceExists()` in HCCM https://github.com/hetznercloud/hcloud-cloud-controller-manager/blob/8775196e5c8423a495333492ecd205dd27d89350/hcloud/instances.go#L74-L84 Which tries to lookup the Server in Hetzner Cloud API by ID or by...

#### v1.19.0 added support for Robot servers in HCCM, so users that only want to have Hetzner Cloud & Robot servers in the same cluster, check out [Clusters with Robot...

There are two issues here: ## `kubernetes/cloud-provider` assumes single (cloud) provider This is an official Kubernetes library that we rely on to talk to Kubernetes. We "only" implement a few...

Hey @nmuntyanov, adding Nodes to the Load Balancer happens asynchronously to the initialization. [By default](https://github.com/kubernetes/cloud-provider/blob/fb1011e79d1e032b02462a2eea25612ff44dd65a/controllers/service/controller.go#L997-L1001) it adds any Nodes that are initialized & marked as `Ready`. Maybe your Node was...

You did not mention which Node you expected to be added in the latest logs, I will assume `pool1-7e9ea74aec95871f` as that one is being initialized there. As we can see...

Hey @guettli, the docs for this annotation are really missing the point. If the annotation is set, the IPv6 address of the LB is _NOT_ written back into the Kubernetes...

Hey @guettli, is the offer to create a PR still standing?

Quick question: Would you need one certificate per domain or are multiple domains in one certificate fine? (Through [Subject Alternative Name](https://en.wikipedia.org/wiki/Subject_Alternative_Name))

> In the interim, I propose configuring the [reconciliation period](https://github.com/kubernetes/cloud-provider/blob/0dac8cc4d471bfbfdf48e7d6a9c039efb8d21d88/config/v1alpha1/defaults.go#L69) from 10s -> 30s. This means that new servers take a little longer to get set up, but the normal...

Unfortunately this is not fixed yet. Reconciliation still happens on a 30s interval (or whatever people configure). This can only be properly fixed by upstream `k/cloud-provider`: https://github.com/kubernetes/kubernetes/issues/60646