Service - Intensity setting should accept both RowLevelRepair and Legacy modes

[fee-mendes]

The default Scylla Manager behavior is to detect whether a cluster supports RowLevelRepair and - if it does - default to it. Otherwise, we fallback to the old default controller (Legacy mode).

This behavior can be overriden via the force_repair_type tunable, but it essentially makes the possible values of choice mutually exclusive when adjusting the --intensity flag.

Latency sensitive workloads will typically prefer a legacy behavior, whereas workloads with defined peak periods will typically prefer to adjust repair accordingly to RowLevelRepair mechanisms.

Scylla Manager should be capable of auto choosing between the two repair controller types depending on the user input:

• A --intensity 0.1 parameter should trigger the Legacy behavior, whereas
• A --intensity N parameter should trigger the RowLevel controller

@vladzcloudius @mmatczuk

[karol-kokoszka]

Closing, intensity 0.1 will be depracated in 3.2 https://github.com/scylladb/scylla-manager/issues/3424

[vladzcloudius]

Closing, intensity 0.1 will be depracated in 3.2 #3424

Let's reconsider. See my comments on #3424