User Story

As a user it would be convenient to have the default container set on OpenShift deployments so that when I get logs from the command line it is easier to know which container I should be inspecting.

Background

Inspired by this comment in slack , it would make our users' lives easier if they didn't need to always specify the container name when getting logs.

For example, this scenario is common:

 $ oc -n openshift-machine-api logs -l k8s-app=cluster-autoscaler-operator --tail 3
Defaulted container "kube-rbac-proxy" out of: kube-rbac-proxy, cluster-autoscaler-operator
I1011 16:26:34.230712       1 main.go:337] Reading certificate files
I1011 16:26:34.230968       1 main.go:371] Starting TCP socket on 0.0.0.0:9192
I1011 16:26:34.231772       1 main.go:378] Listening securely on 0.0.0.0:9192

If the default container were set this operation would have given the logs for the cluster-autoscaler-operator instead of the kube-rbac-proxy.

The Kuberenetes community has defined a  well-known annotation to specify the default container, `kubectl.kubernetes.io/default-container`. This should be added to the machine-api-operator, cluster-autoscaler-operator, and cluster-cloud-controller-manager-operator, as well as all the individual deployments from those repositories.

In addition, for convenience, the first container of every deployment should be specified as the default container, eg the first container should be the important one.

Steps

• Update deployments in MAO, CAO, and CCCMO

Stakeholders

• openshift eng

Definition of Done

• users can get logs using the default container for the primary targets of MAO, CAO, and CCCMO

• Docs

• n/a

• Testing

• n/a