Create ADR on options for an operator to manage stackable components

まず OpenShift Slack の議論と提案されている4つのアプローチを確認し、その後、想定される構成と承認プロセスを把握するために、ドキュメントリポジトリにある既存の ADRs を調査します。OpenShift 認証と、より広範な Stackable operator の選択肢を評価し、ADR が承認されてマージされ、1つの選択肢が選ばれ、フォローアップチケットが作成された時点で作業完了とします。

This was triggered to a certain extent by a discussion around OpenShift and how we cannot currently certify the secret and listener operator, because these need to be deployed a a DaemonSet, which is not supported by Operator Lifecycle Manager.

The best practice / workaround here seems to be to create an operator which in turn deploys the necessary DaemonSets and then certify that operator - which in turns allows us to specify this as a dependency for the rest of our operators.
We had similar discussions a while ago / the topic comes up every now and then, whether we want a stackable operator that is in charge of updating our components / internal configmaps etc. Sort of like a k8s internal stackablectl.

I think it is worth writing an ADR and evaluating the various options that we have to
a) solve the concrete matter of getting secret and listener operator OpenShift certifiable
b) look at this from a bit further away and consider if we need a stackable operator

We briefly discussed this in the openshift slack channel: https://stackable-workspace.slack.com/archives/C03DGUK815E/p1662631718184049 and came up with these options as an initial list (there will probably be more, this is not a conclusive list):

1. Have Stackable operator that works somewhat analogue to stackablectl but from inside the cluster
2. Create an operator specifically for openshift to become compliant, leave everything else as is
3. Rework secret and listener operator minimally to simply deploy one pod which then in turn deploys the daemonset
4. 3 but have that "one pod" be the commons operator

This is done when:

• The ADR is approved and merged and an option was chosen
• Follow up tickets for the chosen option have been created