Have a command for purging orphaned snapshots

[dkropachev]

Currently s-m does not purge snapshots when backup job is deleted In the sake of usability we need to have a way to purge these snapshots in sctool and have it properly documented Another nice to have feature would be an option to delete snapshots when job is deleted

### Tasks
- [ ] https://github.com/scylladb/scylla-manager/issues/3467
- [ ] https://github.com/scylladb/scylla-manager/issues/3468
- [ ] https://github.com/scylladb/scylla-manager/issues/3469

[Michal-Leszczynski]

In the sake of usability we need to have a way to purge these snapshots in sctool and have it properly documented

We have sctool backup delete --cluster <id|name> --snapshot-tag <tag> for deleting specific backups and sctool backup list --cluster <id|name> for listing all backups, so deleting orphaned snapshots can be achieved by combining those commands.

[karol-kokoszka]

All the commands are here. Example:

➜  scylla-manager git:(docs_add_agent_config_change_prereq) ✗ ./sctool backup list --all-clusters -c 364e4cc0-6f69-4f2e-a83a-a35f603d3ac2 -L s3:backuptest-smoke
Cluster: 79eb5d3d-6648-441f-8f10-73fc4aecb5fa

backup/5aca45d0-9ccc-44cb-bf8e-88823621c1ed
Snapshots:
  - sm_20230630133111UTC (2.988G, 3 nodes)
  - sm_20230630133032UTC (2.988G, 3 nodes)
Keyspaces:
  - system_schema (13 tables)
  - backuptest_data (1 table)

Cluster: ea031891-0357-436c-acea-9fb7f1661a90

backup/0dc7c09b-d083-45e8-8c14-ad8ab7ca0f95
Snapshots:
  - sm_20230710220000UTC (4.008M, 6 nodes)
Keyspaces:
  - system_traces (5 tables)
  - system_distributed_everywhere (1 table)
  - system_distributed (4 tables)
  - system_schema (13 tables)
  - system_auth (4 tables)

backup/5c6803ee-9674-4de8-9414-13335eacd2e6
Snapshots:
  - sm_20230710221000UTC (3.488M, 6 nodes)
Keyspaces:
  - system_distributed_everywhere (1 table)
  - system_schema (13 tables)
  - system_auth (4 tables)
  - system_traces (5 tables)
  - system_distributed (4 tables)

https://manager.docs.scylladb.com/stable/sctool/backup.html#backup-list https://manager.docs.scylladb.com/stable/sctool/backup.html#backup-delete

But sctool backup delete requires to pass cluster-id. If the cluster doesn't exist then it fails. Hehehe, so seems it's useless in the current state.

➜  scylla-manager git:(docs_add_agent_config_change_prereq) ✗ ./sctool backup delete -T sm_20230710221000UTC -L s3:backuptest-smoke 
Error: required flag(s) "cluster" not set