Anomalo Data Quality

Anomalo data testing

Anomalo runs rule-based checks on SQL tables, compares tables, shows the rows that fail, and checks unstructured documents for missing metadata, PII and corrupted files.

• SQL Table Tests: Over 40 kinds of no-code validation checks (nulls, zeros, duplicates, uniqueness, ranges and other thresholds), plus custom SQL queries for anything the UI does not cover (website).

• Unstructured Data Tests: Anomalo Unstructured runs 15+ LLM-based checks plus custom prompt-defined checks on documents (missing metadata, PII, corrupted files) and scores each document 1-10. Checks target document stores such as S3, not warehouse text columns (website).

• Table Comparison Tests: A "Compare two aggregate SQL queries" validation rule compares a source and target table, used for cloud migration reconciliation (website).

• Inspect Failed Rows: Provides samples of good and bad rows to diagnose failed checks (website).

• DataFrame Tests: Checks run against connected warehouses and databases; no in-memory DataFrame (Pandas, Spark, Polars) validation library is offered (website).

• Data Diff: Table reconciliation compares aggregate queries and samples of rows; no row-by-row, value-by-value diff of two tables is described (website).

• Data Contracts: No data contract feature (producer-consumer schema and quality SLA agreements) is described on product pages (website).

Anomalo data observability

Anomalo profiles tables and detects anomalies with unsupervised ML that learns what normal data looks like. Monitors are configured in the UI.

Compare: data observability tools · data profiling tools

• Data Profiling: Data profiling visualizations show how values are distributed, with samples and visualizations for deeper analysis (website).

• Anomaly Detection: Unsupervised machine learning learns historical patterns and flags anomalies in data values, columns and segments without hand-set thresholds (website).

• Freshness Monitors: Monitors data freshness across the whole warehouse automatically (not for views or Databricks external tables) (website).

• Volume Monitors: Alerts when updates do not add the expected number of rows, based on historical patterns (website).

• Schema Monitors: Alerts when columns or tables are dropped or changed, and checks that tables and views exist (website).

• Field Monitors: Metrics monitoring tracks column-level statistics over time, and AI checks flag anomalies in specific columns (website).

• Distribution Monitors: Unsupervised anomaly detection finds changes in the data values themselves, including distribution shifts (website).

• Segment Monitors: Monitors metrics for data and its key segments, and pinpoints drops in segment records (website).

• Custom SQL Monitors: Users can enter a custom SQL query at any time to define a check the UI does not support (website).

• UI-Based Monitors: No-code check configuration in the UI, with bulk configuration of tables (website).

• Monitors-as-Code: Checks are configured in the UI or through the API; no YAML or version-controlled monitor definitions are described (website).

• Pipeline Monitoring: Detects pipeline interruptions through their effect on data, but does not track orchestrator job durations or run status (website).

• Partition-Based Monitors: AI checks sample 10,000 records from the most recent data, which implies time-sliced scans, but explicit partition or time-filter configuration is not documented publicly (docs require login) (website).

• Window-Based Monitors: Checks compare a sample of the most recent data against earlier days, but configurable tumbling or sliding windows are not documented publicly (website).

Anomalo incident management

Anomalo triages issues, opens tickets in Jira or ServiceNow, and pages on-call teams through PagerDuty or Opsgenie.

• Alert Notifications: Alerts go to Slack, Microsoft Teams, email, PagerDuty, Opsgenie and webhooks (website).

• Incident Triage: Triage and resolution workflows, with lineage showing affected downstream consumers (website).

• Incident Resolution: Automated root cause analysis shows good and bad sample rows and highlights the most anomalous columns and segments (website).

• Ticketing Integrations: Integrates with Jira and ServiceNow (website).

• On-Call Integrations: Integrates with PagerDuty and Opsgenie (website).

• Alert Feedback: Product pages mention "intelligent false positive suppression" with automatic secondary checks, but no user feedback on alerts is documented publicly (website).

Anomalo data quality analytics

Anomalo's data quality dashboards show results across tables.

• Data Quality Dashboards: An executive dashboard shows data quality trends and repeat offenders (website).

• Data Health Scores: Documents get 1-10 quality scores in Anomalo Unstructured; no health score for tables is described (website).

• Incident Resolution Metrics: Management reporting is listed, but MTTD or MTTR metrics are not described (website).

Anomalo data lineage

Anomalo has table-level lineage with impact analysis, but its lineage stays inside the warehouse.

• Table-Level Lineage: Automated lineage shows tables and how data flows between them, fetched from the warehouse daily (website).

• Impact Analysis: Shows how data quality issues affect downstream consumers, linked from alerts (website).

• End-to-End Lineage: Lineage relies only on warehouse records (Databricks, Snowflake, BigQuery) and excludes steps such as Airflow transformations; BI tools are not included (website).

• Column-Level Lineage: Lineage is shown at the table level; column-level lineage is not mentioned (website).

Anomalo data governance

Anomalo syncs data quality results to catalogs such as Alation, Atlan, Collibra, DataHub and Unity Catalog.

• Data Catalog Integration: Pushes data quality results to Alation, Atlan, Collibra, DataHub, data.world, Unity Catalog, Snowflake Horizon and Microsoft Purview (website).

• Data Catalog: Tracks quality metrics, ownership and change history per table, and a documentation agent writes descriptions, but there is no searchable catalog; results are pushed to external catalogs (website).

• Data Products: No data product feature is described (website).

• Data Classification: Unstructured checks flag sensitive PII in documents, but no classification of warehouse columns is described (website).

Anomalo workflows

Anomalo has a REST API, a Python SDK, webhooks and a dbt integration. It doesn't run checks in CI/CD pipelines.

• Scheduling: Checks run on a schedule (hourly by default for metadata, with configurable frequency), or from Airflow and dbt (website).

• Webhooks: Rich API and webhook support (website).

• REST / GraphQL API: Checks can be configured and run programmatically through Anomalo's API (website).

• Python SDK: The anomalo PyPI package provides a REST API client and command-line client (docs).

• dbt Integration: Built-in integrations with Airflow and dbt to run checks in pipelines (website).

• CI/CD Integrations: No CI/CD or pull request integration is described; pipeline runs go through Airflow and dbt (website).

• MCP Server: No MCP server is listed on the integrations or AIDA pages (website).

Anomalo AI agents

Anomalo has an AI assistant that answers questions in plain language and turns them into SQL. It also writes checks from prompts, generates checks automatically, and uses AI to triage issues.

• AI Chat Assistant: AIDA is a conversational copilot that answers questions in natural language and turns them into SQL (website).

• Natural Language Test Generation: AIDA creates new data quality checks from natural language and pre-fills configurations (website).

• Automated Test Generation: AI checks monitor tables automatically with unsupervised ML, with no rules to write (website).

• AI Issue Triaging: Automated root cause analysis and the Data Insights agent explain what changed, likely causes and affected data (website).

• AI Issue Resolution: Agents explain issues and suggest actions, but no automated fixes are described; the Data Issue First Responder agent is marked coming soon (website).