We should add observability into 'dead'/MVCC garbage data that has not been cleaned up which can potentially be impacting workload performance.

This relates to the following issue but provides a higher-level view at the table level on the Databases page.

One potential user flow: SQL Activity -> Statements -> Details view -> Plan View -> Index details

Slack context.

Column name: % of Non-live Data

Tool tip description: % of total logical data that has not been modified (updated or deleted). A low percentage can cause statements to scan more data (MVCC values) than required, which can reduce performance.

Jira issue: CRDB-16540