問題文
A team must onboard new clusters frequently and wants the set of applications on each one to follow from its attributes. Which arrangement fits?
選択肢
- Have the onboarding process call the delivery system's interface to create the application definitions, so that the set of applications is decided at the moment the cluster is registered rather than being derived from anything stored.
- Declare a generator that selects clusters by their labels and produces one application definition per match.
- Copy the existing application definitions by hand and change the destination field in each copy before applying it.
- Install a separate reconciler in each new cluster and give it its own repository, so that the set of applications on a cluster is whatever that cluster's repository contains and no central definition has to be updated when a cluster is added.