# iceDQ Data Quality

> Data reliability platform for data testing, source-to-target reconciliation, and monitoring, with rules in SQL or Groovy or generated by AI from ETL code, as SaaS or installed in your own environment.

Source: https://toolsfordata.com/tools/icedq/ · Updated 2026-11-13

[Website](https://icedq.com/) · [Docs](https://docs.icedq.com/)

Best for: enterprises testing ETL pipelines and data migrations that need full-volume reconciliation across databases, files, and BI reports.

## Overview

- Core Strength: Full-volume reconciliation — record-level exception reports, with ETL testing and observability on one platform
- Test Types: Partial — SQL, table compare, data diff, failed rows
- Detection: Full — profiling, anomaly detection, pipeline monitoring
- Monitor Coverage: 6 of 11 types — UI
- Incident Workflow: Partial — resolution, ticketing
- Quality Analytics: Partial — dashboards
- Lineage Depth: Partial — impact analysis
- Governance Overlap: Not confirmed
- FinOps & AI Observability: Not confirmed
- AI Agent Features: Partial — NL test gen, auto test gen
- Developer Surface: Partial — API, MCP, dbt, CI/CD…
- Deployment Model: Partial — SaaS, self-hosted
- Security & Access: Partial — SSO, RBAC, audit log, certifications
- Pricing: Custom — no public pricing
- Free Option: Free tier (trial on request)
- Notable Caveat: Docs cover installation and security only; most feature claims rest on product pages

## What it's used for

iceDQ is used by enterprises to test ETL pipelines, data migrations, and BI reports: rules check completeness, ranges, and referential integrity, and reconciliations compare source and target at full volume.

## Pricing

iceDQ does not publish prices; it is licensed per user, per server, or enterprise-wide, with a trial on request.

## Deployment

iceDQ runs as a SaaS managed by iceDQ, or customer-managed in your VPC, on-premises, or air-gapped on Kubernetes.

## Pricing

- Model: per user, per server
- Free: Free (trial on request)
- Entry team price: Custom
- Entry enterprise price: Custom
- Source: https://icedq.com/request-a-demo (checked 2026-11-12)

## Features

### Data Testing

- SQL Table Tests: Yes — Validation rules check nulls and blanks, ranges and thresholds, formats, and referential integrity on tables, with SQL expressions or Groovy for custom logic. ([source](https://icedq.com/product/data-testing-tool?view=full))
- DataFrame Tests: Unknown — No pandas, Spark, or Polars DataFrame validation is described; Spark appears only as a query engine for lakehouse testing. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Unstructured Data Tests: No — JSON and XML files are tested for structure; no LLM or semantic checks on free-text or JSON columns are described. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Table Comparison Tests: Yes — Checksum reconciliation compares sums, counts, averages, min, and max values between source and target, optionally by segment. ([source](https://icedq.com/data-reconciliation-tool))
- Data Diff: Yes — Row and column level reconciliation matches records between source and target, with an exception report of source and target values for each mismatch. ([source](https://icedq.com/data-reconciliation-tool))
- Data Contracts: Unknown — The monitoring page describes checks built from data contracts on schema, format, and delivery timing, but no machine-readable contract format. ([source](https://icedq.com/data-monitoring))
- Inspect Failed Rows: Yes — Each run produces an exception report listing record- and column-level mismatches with the source and target values. ([source](https://icedq.com/product/data-testing-tool?view=full))

### Data Observability

- Data Profiling: Yes — Its website says Autometrics profiles data and generates metrics automatically to scale observability. ([source](https://icedq.com/features?view=full))
- Anomaly Detection: Yes — Anomaly detection combines ML, statistical, and rule-based methods, with thresholds that adapt to seasonality, holidays, and trends. ([source](https://icedq.com/products/data-observability-tool))
- Freshness Monitors: Yes — Observability detects stale data and delayed refreshes against an expected timeframe and alerts when a dataset is late. ([source](https://icedq.com/products/data-observability-tool))
- Volume Monitors: Yes — Observability flags unexpected spikes, drops, or missing records against expected patterns, with threshold-based alerts. ([source](https://icedq.com/products/data-observability-tool))
- Schema Monitors: Yes — Observability detects added, removed, renamed, or retyped columns across databases, files, and APIs. ([source](https://icedq.com/products/data-observability-tool))
- Field Monitors: Unknown — Business KPIs and value distributions are tracked, but per-column statistics as time series are not named. ([source](https://icedq.com/products/data-observability-tool))
- Distribution Monitors: Yes — Data drift monitors detect shifts in distributions, patterns, and value frequencies. ([source](https://icedq.com/products/data-observability-tool))
- Segment Monitors: Yes — Segmented metrics apply metrics such as volume to subgroups of the data, for example volume by country code. ([source](https://icedq.com/features?view=full))
- Custom SQL Monitors: Unknown — SQL expressions are described for test rules, but no SQL metric tracked over time as a monitor. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Partition-Based Monitors: Unknown — Partition-based monitoring is not described; observability covers freshness, volume, schema, and drift. ([source](https://icedq.com/products/data-observability-tool))
- Window-Based Monitors: Unknown — Window-based monitoring is not described; observability uses thresholds and ML detection per dataset. ([source](https://icedq.com/products/data-observability-tool))
- UI-Based Monitors: Yes — Observability checks are configured in the UI, or generated from natural language with a GenAI assistant. ([source](https://icedq.com/products/data-observability-tool))
- Monitors-as-Code: Unknown — Rules and workflows can be exported and imported between environments, but code or YAML definitions in Git are not described. ([source](https://icedq.com/features?view=full))
- Pipeline Monitoring: Yes — Observability detects delays in ETL and ELT jobs and scheduled workflows; job duration statistics are not described. ([source](https://icedq.com/products/data-observability-tool))

### Incident Management

- Alert Notifications: Yes — Alerts go to email, Slack, and Microsoft Teams when a rule or workflow fails or succeeds. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Incident Triage: Unknown — Anomalies can be routed to Incident Management, but owner assignment, priority, and grouping are not described. ([source](https://icedq.com/products/data-observability-tool))
- Alert Feedback: Unknown — Only per-dataset sensitivity tuning is described; marking alerts as false positives or expected is not. ([source](https://icedq.com/products/data-observability-tool))
- Incident Resolution: Yes — Its website says detected anomalies are routed to Incident Management for tracking and resolution; resolution states are not described. ([source](https://icedq.com/products/data-observability-tool))
- Ticketing Integrations: Yes — The integrations page lists Jira and ServiceNow as ticketing integrations. ([source](https://icedq.com/integrations))
- On-Call Integrations: No — The integrations cover test management, ticketing, DevOps, and schedulers; PagerDuty and Opsgenie are not listed. ([source](https://icedq.com/integrations))

### Data Quality Analytics

- Data Quality Dashboards: Yes — Dashboards show pass, fail, and trend metrics for every workflow run across pipelines, domains, and environments. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Data Health Scores: Unknown — No health score is named; dashboards show pass, fail, and trend metrics instead. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Incident Resolution Metrics: Unknown — No MTTD, MTTR, or other resolution-time metrics are described. ([source](https://icedq.com/products/data-observability-tool))

### Data Lineage

- Table-Level Lineage: Unknown — A lineage module is mentioned for root cause and impact tracing, but its granularity is not described. ([source](https://icedq.com/products/data-observability-tool))
- Column-Level Lineage: Unknown — A lineage module is mentioned, but column-level lineage is not. ([source](https://icedq.com/products/data-observability-tool))
- End-to-End Lineage: Unknown — Its website mentions end-to-end lineage to find the layer where an issue started, without describing the lineage itself. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Impact Analysis: Yes — Its website says impact analysis and root cause analysis trace issues through the lineage module, including the reports a schema change affects. ([source](https://icedq.com/products/data-observability-tool))

### Data Governance

- Data Catalog: Unknown — No built-in catalog is described; the security docs say iceDQ stores metadata, credentials, rules, and results. ([source](https://docs.icedq.com/security))
- Data Catalog Integration: No — The integrations page lists no catalog tools such as Collibra, Alation, or Unity Catalog. ([source](https://icedq.com/integrations))
- Data Classification: Unknown — Classifications appear only as a business rule type; automatic PII detection is not described. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Data Products: Unknown — Workflows group rules by pipeline, domain, or release, but data products are not described. ([source](https://icedq.com/product/data-testing-tool?view=full))

### FinOps

- Query Performance: Unknown — No warehouse query metrics are described; performance refers to rule execution speed. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Cost Monitoring: Unknown — No warehouse spend or cost monitoring is described. ([source](https://icedq.com/product/data-testing-tool?view=full))

### LLMOps

- AI Observability: No — AI here means ML anomaly detection on data and rule-generation agents; monitoring LLM outputs is not described. ([source](https://icedq.com/products/data-observability-tool))

### Workflows

- Scheduling: Yes — A built-in scheduler runs rules and workflows on a time-based schedule. ([source](https://icedq.com/features?view=full))
- CI/CD Integrations: Yes — Jenkins, Bamboo, Azure DevOps, and GitHub Actions trigger workflows and get pass or fail exit codes. ([source](https://icedq.com/integrations))
- Circuit Breakers: Unknown — Public docs cover installation; no page describes failed rules stopping ETL jobs. ([source](https://docs.icedq.com/))
- Webhooks: Yes — A webhook integration is listed among the pre-built integrations; its payload is not described. ([source](https://icedq.com/integrations))
- REST / GraphQL API: Yes — A REST API invokes any rule, workflow, or schedule programmatically. ([source](https://icedq.com/product/data-testing-tool?view=full))
- MCP Server: Yes — The security docs describe MCP server tools that AI assistants call, with rules still running in the iceDQ engine. ([source](https://docs.icedq.com/security))
- Python SDK: Unknown — No Python SDK is named; the API reference did not load beyond its title. ([source](https://api.icedq.com/docs/ng-api-design/rhobr7ms9bywm-authentication))
- dbt Integration: Yes — Its website says it integrates with orchestration tools including Airflow, Azure Data Factory, and dbt. ([source](https://icedq.com/data-reconciliation-tool))

### AI Agents

- AI Chat Assistant: Unknown — A GenAI assistant builds checks from natural language, but a chat interface for questions about data health is not described. ([source](https://icedq.com/products/data-observability-tool))
- Natural Language Test Generation: Yes — An AI agent generates data testing rules from plain-English descriptions, ETL code, or mapping documents. ([source](https://icedq.com/product/data-testing-tool?view=full))
- Automated Test Generation: Yes — The Bulk Rule Generator creates reconciliation rules across whole schemas at once. ([source](https://icedq.com/data-reconciliation-tool))
- AI Issue Triaging: Unknown — AI is described for anomaly detection and rule generation, not for triaging issues. ([source](https://icedq.com/products/data-observability-tool))
- AI Issue Resolution: Unknown — No AI remediation is described. ([source](https://icedq.com/products/data-observability-tool))

### Security

- SSO: Yes — SSO works through OpenID Connect with providers such as Okta and Entra ID, plus LDAP. ([source](https://docs.icedq.com/security))
- SCIM Provisioning: No — The security docs cover OIDC SSO, LDAP, and roles, with no SCIM provisioning. ([source](https://docs.icedq.com/security))
- RBAC: Yes — Owner, Contributor, Executor, and Reader roles are assigned at account or workspace level. ([source](https://docs.icedq.com/security))
- Audit Log: Yes — Audit logs record authentication, administrative actions, rule executions, and system events, and can be forwarded to a SIEM. ([source](https://docs.icedq.com/security))
- Compliance Certifications: Yes — The security docs state ISO/IEC 27001 certification and a SOC 2 Type II attestation. ([source](https://docs.icedq.com/security))

### Deployment

- SaaS Deployment: Yes — An iceDQ-managed SaaS option runs the infrastructure, patching, and backups for you. ([source](https://docs.icedq.com/security))
- Hybrid Deployment: Unknown — The docs describe customer-managed and iceDQ-managed SaaS deployments, not a split control and data plane. ([source](https://docs.icedq.com/security))
- Self-Hosted Deployment: Yes — Customer-managed deployments run in your VPC, on-premises, or air-gapped, installed on embedded or existing Kubernetes clusters. ([source](https://docs.icedq.com/security))
- Cloud-Native App: No — Deployment options are SaaS, customer VPC, on-premises, and air-gapped; no Snowflake Native App or Databricks app. ([source](https://docs.icedq.com/security))

### Open Source

- Open Source: No — iceDQ is commercial, licensed per user, per server, or enterprise-wide, with no open-source edition. ([source](https://icedq.com/overview))

## Integrations

23 integrations, each sourced from the vendor docs.

- BI: Power BI, Tableau, Qlik
- Code Repositories: Azure DevOps
- Data Applications: SAP ERP
- Data Storage: Amazon S3
- Data Warehouse: Snowflake, Databricks SQL, Azure Synapse, Oracle DB, Google BigQuery, Amazon Redshift
- Databases: Microsoft SQL Server, PostgreSQL, MySQL
- Incident Management: Jira, ServiceNow
- Notifications: Email, Slack, Microsoft Teams, Webhook
- Open Table Formats: Apache Iceberg
- Streaming: Apache Kafka

## Alternatives

- [RightData DataTrust](https://toolsfordata.com/tools/rightdata-datatrust/): The closest match: no-code source-to-target reconciliation and regression tests in CI/CD, with certification evidence for audits.
- [BiG EVAL](https://toolsfordata.com/tools/big-eval/): Consider BiG EVAL if your warehouse runs on SQL Server, Oracle, Teradata or SAP and tests must stay on-premises: one metadata-driven check runs across thousands of assets.
- [FirstEigen DataBuck](https://toolsfordata.com/tools/firsteigen-databuck/): Consider DataBuck if you'd rather not write reconciliation rules: its AI infers validation rules from the data, then reconciles across legacy and cloud sources.
- [Datafold](https://toolsfordata.com/tools/datafold/): Consider Datafold for a one-off warehouse migration, where value-level diffs between source and target are the whole job and agents help move the code.
- [Qualytics](https://toolsfordata.com/tools/qualytics/): Consider Qualytics if test coverage matters more than reconciliation: it generates checks from profiling and runs on-premise via Kubernetes.
- [DataComPy](https://toolsfordata.com/tools/datacompy/): The free option for reconciliation tests in code: compare two tables by join keys with numeric tolerances and a mismatch report, in Pandas, Polars, Spark or Snowpark.
