Manager integration stability and correctness [Part 1]

[tnozicka]

At this point, the manager integration is quite buggy and sometimes create hot loops. There is a place for improvement for testing and docs as well.

This epic is meant only to fix what we have, not to create a proper integration that respects namespaces and isolation.

Task list

Beta Give feedback

1. Support cron in ScyllaCluster API for Scylla Manager tasks #1739
   kind/feature priority/important-longterm +2
   
2. Deprecate "interval" in ScyllaCluster's manager task API #1796
   kind/feature priority/important-longterm +2
   
3. Scylla manager controller deletes and creates tasks even without hitting conflicts #1752
   kind/bug priority/critical-urgent +2
   
4. Mgmt backup fails with create task: start date of scheduled task cannot be in the past when it is 12+ backup in a row #1683
   kind/bug priority/important-soon +2
   
5. [Flake] Scylla Manager integration [It] should register cluster and sync repair tasks #1694
   kind/flake priority/important-soon +2
   
6. Extend Scylla Manager E2E with backup and manual restore procedure #1671
   kind/feature priority/critical-urgent triage/accepted +3
   
7. Extend Scylla Manager integration e2e coverage with error propagation testing #1907
   kind/feature priority/important-longterm +2
   

[tnozicka]

@rzetelskik please split this into Part 1 (this issue) that is realistic to ship for the targeted release and Part 2 for the rest.

[rzetelskik]

@tnozicka done, #1939
please assign the priority and targeted release