You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If a rollout resource is removed from the cluster the referenced deployment from "workloadRef" keeps the entry "replicas: 0" (if "scaleDown" is used). It will never scale up. This causes a service interruption.
we use this in a helm chart which allows to switch off the use of argo rollouts (and switch back to use the deployment directly). If we switch it off - the rollout resource is removed (by argo-cd). And the deployment stays with "replicas: 0" forever.
We also use a hpa --> it ignores that state with "scaling is disabled since the replica count of the target is zero"
It would be great if the argo-rollouts controller can detect this and remove the "Replicas: 0" if a rollout is removed.
version of argo rollouts: 1.7.2
Message from the maintainers:
Impacted by this bug? Give it a 👍. We prioritize the issues with the most 👍.
The text was updated successfully, but these errors were encountered:
If a rollout resource is removed from the cluster the referenced deployment from "workloadRef" keeps the entry "replicas: 0" (if "scaleDown" is used). It will never scale up. This causes a service interruption.
use Case:
we use "scaleDown" in our rollout spec:
we use this in a helm chart which allows to switch off the use of argo rollouts (and switch back to use the deployment directly). If we switch it off - the rollout resource is removed (by argo-cd). And the deployment stays with "replicas: 0" forever.
We also use a hpa --> it ignores that state with "scaling is disabled since the replica count of the target is zero"
It would be great if the argo-rollouts controller can detect this and remove the "Replicas: 0" if a rollout is removed.
version of argo rollouts: 1.7.2
Message from the maintainers:
Impacted by this bug? Give it a 👍. We prioritize the issues with the most 👍.
The text was updated successfully, but these errors were encountered: