K8SSAND-1821 ⁃ Ability to merge cluster- and dc- level configs

[adutra]

What is missing?

Currently cluster-level configs are coalesced with dc-level ones in an all-or-nothing fashion: if the dc-level config is present, then the cluster-level one is ignored.

Why do we need it?

With #710, YAML config files are now schemaless in the CRD and represented internally as map[string]interface{}. It should be fairly simple to use utils.MergeMapNested to merge cluster-level and dc-level configs.

Note: JVM options are still schemaful so we can't apply this fix to the JvmOptions struct. We might be able to write custom code to handle it, or resort to merging libraries as suggested in #526.

┆Issue is synchronized with this Jira Task by Unito
┆friendlyId: K8SSAND-1821
┆priority: Medium

[adutra]

This is going to be superseded by #525, I have a PR up for that: #780.