WebRedshift Health

Table Analysis

The Table Analysis dashboard provides deep-dive metrics for every table in the connected Redshift cluster. It surfaces vacuum health, encoding efficiency, and data quality signals in a searchable, sortable table with drill-down detail panels.

Table Metrics Table

The main table lists all Redshift user tables with the following columns:

Column	Source	Description
Table Name	svv_table_info	Schema-qualified table name
Size (MB)	svv_table_info	Total on-disk size across all nodes
Row Count	svv_table_info	Estimated visible row count (excludes ghost rows)
Sort Key	svv_table_info	First sort key column name
Dist Style	svv_table_info	Distribution style: KEY, EVEN, ALL, or AUTO
Last ANALYZE	stl_analyze	Timestamp of most recent ANALYZE operation
Last VACUUM	stl_vacuum	Timestamp of most recent VACUUM operation
Unsorted %	svv_table_info	Percentage of rows not in sort key order. High values degrade query scan performance.
Ghost Row %	svv_table_info	Percentage of rows marked for deletion but not yet reclaimed by VACUUM
Alert Flags	stl_alert_event_log	Icons for active alert types: sort key, deletes, NL join, distribution, broadcast, stale stats

Health Indicators

Row-level color coding highlights tables requiring attention:

Red: Unsorted % > 20% or Ghost Row % > 30% — immediate VACUUM recommended

Yellow: Last ANALYZE > 7 days or Last VACUUM > 14 days — maintenance due soon

Green: All metrics within healthy thresholds

Table Detail Panel

Clicking a table row opens a slide-in detail panel with six sections:

Column Encoding — Each column with its compression encoding type and selectivity. Columns with raw encoding on large tables are flagged as optimization opportunities.
Sort Key Effectiveness — Ratio of zone maps skipped vs. scanned, indicating how well the sort key aligns with query predicates.
Vacuum History — Timeline of past VACUUM runs showing duration and rows reclaimed per run.
Alert Event Log — Recent alert events from stl_alert_event_log for this table with timestamps and event descriptions.
DML Volume — Inserts and deletes since last ANALYZE, sourced from stl_insert and stl_delete.

Ghost Row Formula

Ghost Row % = (tbl_rows − estimated_visible_rows) / NULLIF(tbl_rows, 1) × 100. Ghost rows consume storage and slow down queries. Run VACUUM DELETE ONLY to reclaim space without resorting.