Discovery scans your databases to build a catalog of tables, views, and columns. Run it manually or on a schedule to keep your asset inventory current.Documentation Index
Fetch the complete documentation index at: https://docs.anomalyarmor.ai/llms.txt
Use this file to discover all available pages before exploring further.
Running Discovery Manually
- Go to Data Sources in the sidebar
- Find your data source
- Click Run Discovery (or the refresh icon)
- Wait for completion (typically 30 seconds to 5 minutes)
Scheduling Discovery
Set discovery to run automatically:- Go to Data Sources → select your data source
- Click Settings (gear icon)
- Under Discovery Schedule, select frequency:
| Schedule | Best For | Typical Use Case |
|---|---|---|
| Hourly | Production databases | Catch changes quickly |
| Every 6 hours | Active development | Balance coverage and load |
| Daily | Stable databases | Low-change environments |
| Weekly | Archives, warehouses | Rarely modified systems |
- Click Save
What Discovery Finds
During each run, discovery catalogs:| Asset Type | What’s Captured |
|---|---|
| Tables | Name, schema, columns, types, constraints |
| Views | Name, schema, underlying columns |
| Columns | Name, data type, nullability, defaults |
| Relationships | Foreign keys, primary keys |
Discovery Duration
Discovery time depends on database size:| Database Size | Typical Duration |
|---|---|
| < 100 tables | 30 seconds |
| 100-500 tables | 1-2 minutes |
| 500-2000 tables | 2-5 minutes |
| 2000+ tables | 5-15 minutes |
Discovery runs read-only queries against system catalogs. It doesn’t impact your database performance significantly.
Incremental vs Full Discovery
Incremental (default): Checks for changes since last run. Faster. Full: Re-scans everything. Use when:- First run after connecting
- Suspect missed changes
- After major database migrations
- Go to Data Sources → select your data source
- Click the dropdown arrow next to Run Discovery
- Select Full Discovery
Viewing Discovery Results
After discovery completes:- Go to Assets in the sidebar
- Filter by data source to see discovered assets
- Click any asset to view details
Discovery History
View past discovery runs:- Go to Data Sources → select your data source
- Click History tab
- See each run with:
- Timestamp
- Duration
- Assets found/changed
- Any errors
Troubleshooting
Discovery stuck or taking too long
Discovery stuck or taking too long
Causes:
- Large database with many tables
- Slow network connection
- Database under heavy load
- Check discovery history for progress
- Wait up to 30 minutes for large databases
- If stuck, cancel and retry during off-peak hours
Discovery completed but missing tables
Discovery completed but missing tables
Causes:
- User doesn’t have permission on those schemas
- Tables in schemas excluded from discovery
- Views the user can’t access
- Check your database user’s permissions
- Grant
SELECToninformation_schema(or equivalent) - Verify schema inclusion settings
Discovery failed
Discovery failed
Causes:
- Connection credentials changed
- Database unreachable
- Permission revoked
- Check the error message in discovery history
- Test the connection in Data Sources
- Verify credentials and network access
Common Questions
Does discovery impact database performance?
No. Discovery runs read-only queries against system catalogs (information_schema and equivalents), not against your data tables. The load is negligible even on large databases.
How often should I run discovery?
For production databases, run discovery hourly to catch schema changes quickly. Stable databases can run daily, and archives can run weekly. More frequent runs shorten the window between a change and detection.What’s the difference between incremental and full discovery?
Incremental discovery checks only for changes since the last run and is the default. Full discovery re-scans everything and should be used after the first connection, after major migrations, or when you suspect a missed change.Why are some tables missing after discovery?
The most common cause is permissions. The database user AnomalyArmor connects with needsSELECT on information_schema (or the vendor equivalent) and access to the target schemas. Check also that the schema isn’t excluded in your discovery settings.
How long does discovery take on a large database?
Databases under 100 tables finish in about 30 seconds. 500-2000 tables take 2-5 minutes, and 2000+ tables can take 5-15 minutes. If a scan runs longer than 30 minutes it’s likely stuck, retry during off-peak hours.See Also
Discovery Concepts
How discovery works under the hood
Schema Changes
Understanding detected changes