ketall
ketall copied to clipboard
Add separator between resource types
Is your feature request related to a problem? Please describe. Currently we get one list for all resource types. Would be useful to have it grouped by resource type.
Currently
configmap/coredns kube-system 27d
configmap/extension-apiserver-authentication kube-system 27d
configmap/kube-proxy kube-system 27d
configmap/kubeadm-config kube-system 27d
configmap/kubelet-config-1.13 kube-system 27d
endpoints/kube-controller-manager kube-system 27d
endpoints/kube-dns kube-system 27d
endpoints/kube-scheduler kube-system 27d
Describe the solution you'd like Proposal
configmaps:
configmap/coredns kube-system 27d
configmap/extension-apiserver-authentication kube-system 27d
configmap/kube-proxy kube-system 27d
configmap/kubeadm-config kube-system 27d
configmap/kubelet-config-1.13 kube-system 27d
endpoints:
endpoints/kube-controller-manager kube-system 27d
endpoints/kube-dns kube-system 27d
endpoints/kube-scheduler kube-system 27d
Describe alternatives you've considered
I was basing how k get all
works.
Additional context
In case of namespace with a lot of resources - there is easier visually to look at results
I do not insist on specific format, I would like something which clearly distinct different resource types
@grzesuav The long-term plan is to output exactly the same as k get all
. However, the resource printer code is not usable by plugins so far. There is an open issue in kubernetes which will solve this in the future. So expect a major improvement in that regard at some point.
However, until then I can improve the current printer. I'm a bit reluctant to add these headings, as this is also not the way how k get all
does it. Would simply an empty line work for you? This would already separate resource groups visually.
Yes, definitely will be better. Could you also link here upstream issue ? Thank you :)
@grzesuav https://github.com/kubernetes/kubernetes/issues/74179
@grzesuav After trying this out, I don't think that empty lines will improve the readability of the output. For example,
$ ketall -n tiller
NAME NAMESPACE AGE
endpoints/tiller-deploy tiller 17d
pod/tiller-deploy-777d84c777-njlx9 tiller 17d
secret/default-token-skk9k tiller 17d
secret/tiller-user-token-hbqdr tiller 17d
serviceaccount/default tiller 17d
serviceaccount/tiller-user tiller 17d
service/tiller-deploy tiller 17d
deployment.apps/tiller-deploy tiller 17d
replicaset.apps/tiller-deploy-777d84c777 tiller 17d
deployment.extensions/tiller-deploy tiller 17d
replicaset.extensions/tiller-deploy-777d84c777 tiller 17d
Which looks quite odd and does not improve readability for me. This would not change, if the header were repeated for every group.
Besides, I think that no lines in between is more in line with kubectl
, for example look at the output of k get all -oname
. This also prints the full resource name and has no interspersed blank lines, just as ketall
.
However, I'm considering to colorize the result. That might help a bit.
Yes, but I was basing on just k get all
- indeed, adding -oname
looks awful. Hope coloring will help.