Operator reliably creates zombie namespace about prometheus-kubernetes HOT 6 CLOSED

Try kubectl get all -n monitoring then erase all the objects there. Then, kubectl delete ns monitoring --grace-period=0 --force

from prometheus-kubernetes.

Tried that, also tired deleting the controller pod. Doesn't helo. Seems it is this bug.

kubernetes/kubernetes#37278

from prometheus-kubernetes.

try this:

kubectl delete servicemonitor $(kubectl get servicemonitor -n monitoring --no-headers | cut -d " " -f1) -n monitoring
kubectl delete prometheus $(kubectl get prometheus -n monitoring --no-headers | cut -d " " -f1) -n monitoring
kubectl delete alertmanager $(kubectl get alertmanager -n monitoring --no-headers | cut -d " " -f1) -n monitoring

from prometheus-kubernetes.

kubectl delete servicemonitor $(kubectl get servicemonitor -n monitoring --no-headers | cut -d " " -f1) -n monitoring
No resources found.

kubectl delete prometheus $(kubectl get prometheus -n monitoring --no-headers | cut -d " " -f1) -n monitoring
No resources found.

kubectl delete alertmanager $(kubectl get alertmanager -n monitoring --no-headers | cut -d " " -f1) -n monitoring
the server doesn't have a resource type "alertmanager"
error: resource(s) were provided, but no name, label selector, or --all flag specified

Btw, namespace is empty

from prometheus-kubernetes.

all three puts error: resource(s) were provided, but no name, label selector, or --all flag specified

from prometheus-kubernetes.

Please contact me on http://slack.kubernetes.io I'm @camil

from prometheus-kubernetes.