When looking at the "Workloads" within the GKE GUI, it shows some of the deployments as shutdown even though they are working correctly.
When looking at the Pods of such a "shutdown" deployment, there is lots of shutdown pods but one is running (as expected - there should be just one pod running and the others were shutdown for any reason in the past).
I get the same result when using kubectl. The deployments that are marked "OK" have just a one pod so far so they will become "shutdown" in the future as the pods will be shutdown and restarted.
The question is how to get rid of such pods automatically so that it's not confusing and also to shorten the pods listing via kubectl without any additional filtering.
EDIT:
There are deployments (picture 1) managing pods, detail of one of the "shutdown" deployments is in picture 2, listing its pods. The deployment is actually working but is marked as "Shutdown" in the GUI.
kubectl get deploy is giving me the deployments correctly (some deployments have a single replica, some have 2 replicas, that's OK):
kubectl get pods is giving me something like this (all the listed pods belong to a single deployment but there should be just a single pod - the one that is running):
The question remains - how to make these pods to disappear?
EDIT 2:
I found out there is default revisionHistoryLimit set to 10, i.e. there should be 10 past ReplicaSets kept (which seems to be true in the GUI). But even within each (old) ReplicaSet there are multiple pods with the "Shutdown" status. All of this is probably causing the Deployment seems to be in a Shutdown state even though it's working correctly.