Add DynamoDB capacity mode, backup, and DAX rules

[dannysteenman]

What

Three rules extending the dynamodb module: on-demand tables with steady traffic that fit provisioned capacity, aging on-demand backups, and idle DAX clusters.

Why

The existing four DynamoDB rules cover provisioned-mode waste (stale/unused tables, autoscaling gaps) but say nothing about the capacity-mode choice itself — on-demand at steady sustained load costs several times equivalent provisioned throughput. Old on-demand backups and idle DAX clusters are the same "accumulates silently" shapes already flagged elsewhere in the catalog.

Details

1. On-demand table whose consumed capacity is steady and predictable over the lookback window → provisioned would be cheaper — discovery; review severity since the switch has operational implications. (The inverse — provisioned with spiky/idle usage — is partially covered by DYNAMODB-2/-3; define the boundary during research.)
2. On-demand backups older than a review threshold — discovery; direct analog of CLDBRN-AWS-RDS-10 (manual snapshot age). PITR remains exempt.
3. DAX cluster idle — no requests over the lookback window — discovery; analog of CLDBRN-AWS-ELASTICACHE-2. Research: whether DAX fits in dynamodb/ or its own module for the service-name/ID convention.

• Continues CLDBRN-AWS-DYNAMODB-N sequence; hydration datasets follow the existing DynamoDB dataset wiring.