Skip to main content

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.

AnomalyArmor is a data observability platform that helps data teams detect issues before they become incidents. Know when schemas change, data goes stale, or pipelines break, before your stakeholders do.
Data sources flow through AnomalyArmor monitoring (schema, freshness, quality) to alert destinations (Slack, PagerDuty, Email)

​
The Problem We Solve

Data teams face a common challenge: you don’t know what you don’t know.
  • A column gets dropped upstream → your dbt models fail at 3am
  • An ETL job silently stops → dashboards show stale data for hours
  • A data type changes → aggregations return wrong results for days
  • A new table appears → you find out when someone asks about it
These aren’t edge cases. They’re Tuesday. AnomalyArmor gives you visibility so you can catch issues when they happen, not when your CFO asks why the numbers are wrong.

​
What is Data Observability?

Data observability brings the principles of application monitoring to your data infrastructure. Just as you wouldn’t run a production application without alerts for errors and latency, you shouldn’t run data pipelines without visibility into data health.
Comparison showing how AnomalyArmor catches schema changes in minutes versus hours of silent failure without observability

​
Key Features

Schema Monitoring

Automatically discover tables and detect schema changes before they break downstream systems. Know immediately when columns are added, removed, or modified.

Freshness Monitoring

Set SLAs on data freshness and get alerted when data is stale or delayed. Never show yesterday’s data on today’s dashboard.

Data Quality Metrics

Track row counts, null percentages, distinct values, and more. Detect anomalies automatically using statistical analysis.

Intelligent Alerts

Route alerts to Slack, email, PagerDuty, or webhooks. Configure rules to filter noise and escalate what matters.

AI-Powered Insights

Get AI-generated documentation for your tables and columns. Ask natural language questions about your data catalog.

Report Badges

Embed data quality status in Databricks, Looker, Notion, and other tools. See pass/fail at a glance.

​
Who Uses AnomalyArmor?

Your challenge: Pipelines break when upstream schemas change. You’re the first call when data is wrong.How AnomalyArmor helps:
  • Get alerted to schema changes before your pipelines run
  • Understand what changed and when
  • Reduce time-to-resolution for data incidents

​
Supported Data Sources

AnomalyArmor connects to your existing data infrastructure with no agents to install:
Data SourceSupport LevelKey Features
PostgreSQLFullTables, views, schemas, constraints
DatabricksFullUnity Catalog, Delta tables, views
ClickHouseFullTables, materialized views, dictionaries
More data sources coming soon. Contact us if you need a specific connector.

​
How It Works

AnomalyArmor follows a simple, secure model:
  1. Connect: Add your database with read-only credentials
  2. Discover: We scan your metadata (never your actual data)
  3. Monitor: Scheduled scans detect changes automatically
  4. Alert: Get notified through your preferred channels

How AnomalyArmor Works

Deep dive into the architecture and data flow

​
Security First

AnomalyArmor is built with security as a foundation:
  • Metadata only: We never access your actual data
  • Encrypted everywhere: TLS 1.3 in transit, AES-256 at rest
  • Minimal access: Read-only, dedicated database users recommended

Security Overview

Learn about our security practices and compliance

​
Getting Started

Ready to try AnomalyArmor? You can be monitoring your first database in under 15 minutes.

Quickstart Guide

Step-by-step guide to connect your first database and set up monitoring

Core Concepts

Understand assets, discovery, schema changes, and freshness

​
Common Questions

​
Does AnomalyArmor access my actual data?

No. AnomalyArmor only queries metadata (system catalogs like information_schema and pg_catalog) and bounded aggregates (COUNT, MIN, MAX, AVG, distinct counts). It never runs SELECT * or reads raw row values. The open-source Query Gateway enforces this on every query.

​
What databases does AnomalyArmor support?

PostgreSQL (12+), Databricks (with Unity Catalog), and ClickHouse are fully supported today. BigQuery, Snowflake, Redshift, MySQL, and SQL Server are in active development. See Data Sources for the current list and contact us if you need a specific connector.

​
How long does it take to set up AnomalyArmor?

About 15-20 minutes to connect your first database and get your first alert. The quickstart walks through the four steps: connect, discover, alert, metric.

​
Do I need to install anything in my infrastructure?

No. AnomalyArmor is fully managed SaaS. Connections go outbound from our cloud to your database using the credentials you provide. Nothing runs on your side.

​
Will AnomalyArmor slow down my production database?

Impact is minimal. Queries hit only system tables and run lightweight aggregates on a schedule you control. Typical discovery completes in under 60 seconds for up to 1,000 tables. You can point AnomalyArmor at a read replica if you want zero impact on the primary.

​
Can I try AnomalyArmor without talking to sales?

Trials are available by invitation. If you received a cold email or a trial code, use the link provided. Otherwise, start from the homepage try-it form or email hello@anomalyarmor.ai. Trials are 14 days with no credit card; there is no self-serve trial without a code and no free tier.

​
Need Help?

Documentation

You’re here! Browse the sidebar for detailed guides.

Support

Contact our team for help with your account.

FAQ

Answers to common questions.