1. Proposed title of this feature request

Continue with nodes upgrade if a node cannot be drained

2. What is the nature and description of the request?

Currently, if a node cannot be drained for any reason (like a PDB) during an upgrade (originated by both configuration or version upgrade), the upgrade process of the MCP gets blocked.
Customer wants that the upgrade of that node fails, and the process continue with the next node, letting the failing node without upgrade but working with the older configuration.

3. Why does the customer need this? (List the business requirements here)

Customer have applications that only are able to have 1 replica, and they want to control when those applications were stopped. For that, they have PDB with `MaxUnavailable: 0`. As if they kill the application when the MCO tries to upgrade the node, the application can be deployed in a non upgraded node, the application will need to be stopped again when the second node starts the upgrade.

Customer want that the upgrade proccess continue with other nodes without that kind of applications, and stop the applications only at the end of the upgrade (when most of the nodes were already upgraded).

Also, it is not enough to have `maxUnavailable` set to a greater number, as the number of pods in this situation is not deterministic, and several nodes can be stuck at the same time.

4. List any affected packages or components.

MCO