Currently harbor runs on a single VM, with no high availability setup nor disaster recovery.
This task is to define which one we want (or if we want both) and write down the appropriate tasks for them.
Currently harbor runs on a single VM, with no high availability setup nor disaster recovery.
This task is to define which one we want (or if we want both) and write down the appropriate tasks for them.
Not really, it's simpler than that, it's about having some availability with the current setup.
About moving to helm deployment, we would have to consider also if using the same k8s cluster than toolforge or a different one, as that might lead to a chicken-and-egg problem (specially if we start hosting k8s-specific images in the repo).