Skip to main content

General

AnomalyArmor currently supports:
  • PostgreSQL (12+) - including RDS, Aurora, Supabase, and self-hosted
  • Databricks - with Unity Catalog
  • ClickHouse (21.8+) - including ClickHouse Cloud
More databases coming soon. Contact us if you need a specific database.
No. AnomalyArmor only queries metadata from system tables like information_schema. We see:
  • Table and column names
  • Data types
  • Constraints and indexes
  • Timestamps (for freshness monitoring)
We never see row data, actual values, or query your business tables.
Discovery queries your database’s system catalogs to find tables, views, and columns. For PostgreSQL, we query information_schema and pg_catalog. For Databricks, we use Unity Catalog APIs. Each discovery compares against the previous state to detect changes.
EnvironmentRecommended
ProductionHourly
StagingEvery 6 hours
DevelopmentDaily
You can also run discovery on-demand anytime.
Minimal. Discovery runs lightweight queries against system tables, not your data. Typical impact:
  • 3-10 queries per discovery
  • < 1 second per query
  • < 5 seconds total connection time
  • No table scans
For large databases, you can use a read replica.

Schema Monitoring

  • Columns added
  • Columns removed
  • Column types changed
  • Tables added
  • Tables removed
  • Constraints modified
Changes are detected on the next discovery run after they occur.
Not directly. A rename appears as a column removed + column added. You’ll see both events and can correlate them manually. We’re working on automatic rename detection.
Check these in order:
  1. Did discovery run after the change occurred?
  2. Is the table in the monitored schema?
  3. Does the alert rule scope include this asset?
  4. Is the alert rule enabled?
  5. Is the destination configured correctly?

Freshness Monitoring

You specify a timestamp column (like created_at or updated_at). AnomalyArmor queries MAX(column) to find the latest timestamp. If that timestamp is older than your SLA threshold, we alert you.
Without a timestamp column, we can’t monitor freshness. Options:
  • Add a timestamp column to your ETL
  • Use an _etl_loaded_at column if available
  • Skip freshness monitoring for that table
Common causes:
  1. SLA too tight: Loosen the threshold
  2. Wrong column: Use updated_at instead of created_at for tables with updates
  3. Timezone issues: Ensure consistent timezone handling
  4. Weekend/holiday: Data legitimately doesn’t update

Alerts

  1. Start with 5-10 critical tables only
  2. Route by event type (breaking changes → PagerDuty, additions → email)
  3. Exclude dev/test environments
  4. Set lenient thresholds and tighten over time
See Alert Best Practices for details.
Yes. Go to Alerts → Rules, find the rule, and toggle it OFF. The rule won’t fire until you toggle it back ON.
Go to Alerts → Destinations, select your destination, and click Send Test Alert. This sends a test notification to verify the connection.
Yes. When creating an alert rule, you can select multiple destinations. For example, send to both Slack and email.

Security

Database credentials are encrypted at rest using AES-256. Encryption keys are managed by AWS KMS and never exposed. Not even our support team can see your credentials.
  • GDPR compliant (DPA available)
  • HIPAA-ready (BAA available)
Contact security@anomalyarmor.ai for compliance documentation.
View our current IP addresses in Settings → Security in your dashboard. We use static IPs that you can allowlist in your firewall.
Yes, and we recommend it. AnomalyArmor only needs SELECT access to system tables. See the PostgreSQL or other database guides for exact permissions.

Billing & Plans

Yes. All new accounts start with a free trial that includes full access to all features. No credit card required to start.
Usage is based on the number of monitored assets (tables and views). Check your current usage in Settings → Billing.
Yes. Upgrade or downgrade anytime in Settings → Billing. Changes take effect immediately.
You’ll receive a notification and can either upgrade your plan or reduce the number of monitored assets. We won’t cut off service without warning.

Technical

Yes. API documentation is available for Enterprise customers. Contact us for API access.
Yes, through:
  • Webhooks: Send alerts to any endpoint
  • Slack: Native integration
  • PagerDuty: Native integration
  • Email: Any email address
  • API: For custom integrations (Enterprise)
Modern versions of:
  • Chrome (recommended)
  • Firefox
  • Safari
  • Edge
We don’t support Internet Explorer.
Self-hosted deployments are available for Enterprise customers with specific compliance requirements. Contact sales for details.

Troubleshooting

See our Connection Issues guide for step-by-step troubleshooting.
Common causes:
  1. User lacks permissions on information_schema
  2. Schema filter excludes all schemas
  3. Connected to wrong database
  4. No tables exist in monitored schemas
Check:
  1. Alert rule is enabled
  2. Destination is configured correctly (test it)
  3. Rule scope matches the asset
  4. Discovery is running and detecting changes

Still Have Questions?

Contact Support

Get help from our team

Documentation

Browse the full documentation