Expose and provide upcoming "compactions backlog" metric to monitoring

[mkeeneyj]

scylla_compaction_manager_backlog and scylla_compaction_manager_normalized_backlog should be available as a gauge metric in monitoring

This helps us determine when compaction is behind (for example when determining root cause of space usage growth)

[mkeeneyj]

Per scylladb/scylladb#26287, these exist when dynamic shares are used, currently, and will be implemented for static shares

This request is not pending on 26287, just a note

[amnonh]

@mkeeneyj, please provide the ScyllaDB version those metrics will be available. Please provide more details on how these metrics will be used.

[amnonh]

@mkeeneyj ping