-
Notifications
You must be signed in to change notification settings - Fork 532
kubefed: remove FederatedIngress feature #1284
Comments
I'm not familiar with this feature. Which third-party tools we are talking about here? |
@hectorj2f This feature is already behind a feature flag and if we don't want to support it in future, it can always remain so, that is in an alpha state. If we foresee that this can burden us as maintenance debt, I would suggest to remove it when we reach that point (for example failing/flaky/broken tests which none of us wants to fix). If there are other more pressing issues/problems this is causing us now, I am quite curious to know. |
I did miss saying that if it is enabled by default now, we can certainly change that to keep it disbled in the default deploy configuration. |
One of the intentions of this issue was to call the attention of any potential user of this feature. Based on the current reactions, I cannot see anyone relying on this feature at this moment and finding this action as a blocker to use kubefed. With that said, I am happy to follow a specific procedure such as disabling it first, and then delete it at some point in the future. @RainbowMango I don't think there is any user using kubefed to just manage DNS records of Kubernetes Ingress resources created in other clusters. For sure, I can provide more context on which components can satisfy this similar scenario. I'd personally search for DNS discovery services or gateway-connected services for a multi-clusters env. I realize we should provide a more detailed information that describes which steps to follow to achieve the same result using these tools. @irfanurrehman It sounds good to me to disable now, as part of a short term solution, and then delete it in the future. The main reason is to keep the source code focused on the federation and management of resources instead of having extra features that aim to address some networking gaps. This could also be a blocker to add new features to kubefed while having to support these ones. |
LGTM for disabling and eventually remove from the code. This was an experimental feature and i don't know anybody using it. I agree that its a burden to keep maintaining the feature in alpha state. |
I‘m using this feature for muti-cluster dns discovery with external-dns and coredns. |
@yuswift As a short term solution, we would disable it by default. But I'll share my thoughts during the next sig meeting about these features and their possible future. |
Issues go stale after 90d of inactivity. If this issue is safe to close now please do so with Send feedback to sig-testing, kubernetes/test-infra and/or fejta. |
/remove-lifecycle stale |
with latest version, or v0.6.1, is this feature still available? In another word, user guide (https://github.com/kubernetes-sigs/kubefed/blob/master/docs/ingress-service-dns-with-coredns.md) is unavailable, is that right? |
This feature has been disabled by default within v0.5.0 and will be removed in the future(maybe v0.6.2). If you want to try the feature, use v.0.4.1 or before. |
@conquerorAlex Yes, @yuswift is right about the plans for this feature. Are you still experiencing the issue with v0.5.0? Did you open any issue about it ? |
Hi @hectorj2f I think we need to reopen this issue cause we can still create |
Kubefed v2 should be only focused on the federation of resources across kubefed clusters. In the Kubernetes ecosystem there are other third party tools (such as service meshes) that already provide support to the DNS based federated ingress.
As a consequence, kubefed won't create anymore IngressDNSRecord objects https://github.com/kubernetes-sigs/kubefed/blob/master/docs/ingressdns-with-externaldns.md).
Why is this needed:
Therefore, we need to remove any related work as part of kubefed to keep the source code clean and focused on satisfying the federation of resources.
/kind feature
The text was updated successfully, but these errors were encountered: