Kubernetes module add rollout tasks

[Theprim0]

SUMMARY

It would be interesting to include tasks into kubernetes.core to interact with ROLLOUT.

ISSUE TYPE

• Feature Idea

COMPONENT NAME

kubectl rollout

rollout history, rollout restart, etc...

ADDITIONAL INFORMATION

Tasks as we would do with CLI:
kubectl rollout status Deployment/mi-deployment

[gravesm]

Thanks for filing a feature request. Could you be a bit more specific about what you would like to be able to do? I believe the collection should already support the functionality provided by kubectl rollout. The information from status and history should be available from the k8s_info module, but in a machine actionable format. The k8s_rollback module does what undo does. pause and resume can both be accomplished with a simple k8s_json_patch operation to set the /spec/paused property. The only thing that can't be done easily with the collection is the rolling restart.

[MarceloFCandido]

Thanks for filing a feature request. Could you be a bit more specific about what you would like to be able to do? I believe the collection should already support the functionality provided by kubectl rollout. The information from status and history should be available from the k8s_info module, but in a machine actionable format. The k8s_rollback module does what undo does. pause and resume can both be accomplished with a simple k8s_json_patch operation to set the /spec/paused property. The only thing that can't be done easily with the collection is the rolling restart.

@gravesm, which way do you suggest to do do rollout restart with the collection? I would like to apply the rollout restart that can be seen here but without using the ansible.builtin.shell module.

[Theprim0]

@gravesm exactly, i see k8s_rollback module exists, but I mean "rollout". As @MarceloFCandido says, we can't acomplish "rollout restart" with kubernetes.core module neither with k8s_rollback module.

[jdavis-politico]

Came looking for this too. Glad it's being looked into. Thanks.

[grsiepka]

+1 for this feature, but wonder if in the meantime, could you use the kubernetes.core.k8s_scale module to scale the deployment to 0, then back to what it was?

[JossWhittle]

The following works for anyone else who finds this thread.

kubectl rollout status -w deployment/foo-bar -n foo

becomes...

- name: "k8s wait for deployment/foo-bar rollout"
  kubernetes.core.k8s_info:
    kind: Rollout
    namespace: "foo"
    name: "deployment/foo-bar"
    wait: true
    wait_condition:
      type: Succeeded

Maybe this could be added to the examples? https://docs.ansible.com/ansible/latest/collections/kubernetes/core/k8s_info_module.html#examples

[lordofthejars]

+1 to this feature. I need to restart a deployment because of a change in a ConfigMap.

[besmirzanaj]

+1 for this. seems a very useful feature

[ShubhamRasal]

+1 for this. seems a very useful feature

[vcazsdk]

+1 to this feature

[psyntium]

+1

[mohag]

kubectl rollout status -w deployment/foo-bar -n foo

becomes...

- name: "k8s wait for deployment/foo-bar rollout"
  kubernetes.core.k8s_info:
    kind: Rollout
    namespace: "foo"
    name: "deployment/foo-bar"
    wait: true
    wait_condition:
      type: Succeeded

That gives me a (green) output saying:

Failed to find API for resource with apiVersion "v1" and kind "Rollout"

[dyegoe]

+1

I was looking into restarting ArgoCD deployment after patching the configmap

But, meanwhile, I found a way to restart a deployment:

- name: Restart ArgoCD repo-server
  kubernetes.core.k8s:
    state: patched
    kind: Deployment
    name: argocd-repo-server
    namespace: "{{ argocd_namespace }}"
    definition:
      spec:
        template:
          metadata:
            annotations:
              kubectl.kubernetes.io/restartedAt: "{{ now(fmt='%Y-%m-%dT%H:%M:%SZ') }}"

[owwweiha]

+1