TheDocumentation Index
Fetch the complete documentation index at: https://docs.anomalyarmor.ai/llms.txt
Use this file to discover all available pages before exploring further.
/armor:recommend skill analyzes your data assets and suggests what to monitor, which tables are most critical, and what thresholds to use.
Usage
- “What should I monitor?”
- “Suggest tables to add freshness monitoring”
- “What are good thresholds for my orders table?”
- “Which tables are most critical to monitor?”
- “What’s missing from my monitoring setup?”
What It Does
The recommend skill uses three recommendation engines:Freshness Recommendations
Analyzes table update patterns and suggests monitoring intervals:- Identifies tables that update regularly but aren’t monitored
- Suggests check intervals based on historical update frequency
- Prioritizes tables by importance (row count, naming patterns, downstream dependents)
Metric Recommendations
Scans column types and patterns to suggest data quality metrics:- Identifies columns that should have null checks
- Suggests row count monitoring for high-volume tables
- Recommends format validation for email, phone, and ID columns
Coverage Recommendations
Provides a full view of monitoring gaps:- Shows your coverage score and current tier
- Lists unmonitored tables ranked by importance
- Suggests which monitoring types to add for each table
Example
Common Questions
How does the recommend skill decide which tables are ‘most critical’?
It ranks tables by row count, naming patterns (fact/gold tables rank higher than staging), downstream dependents from lineage, and existing tags. Tables with many consumers and no current monitoring surface first.Do the suggested thresholds actually fit my data, or are they generic?
They’re computed from each table’s own update history, not defaults. The skill looks at recent gaps between updates and picks a threshold that would not have fired during normal variation. You can dry-run any suggestion with/armor:test before enabling.
