Also known as: Databricks Lakehouse Monitoring.
At a glance
| Supported data platforms | Databricks only: Unity Catalog Delta tables and managed Iceberg tables |
| Anomaly detection | Yes, freshness and completeness per table; views and foreign tables not supported |
| Freshness and volume monitoring | Yes: commit freshness and 24-hour row counts |
| Schema checks | Not documented as a schema change monitor |
| Validation rules | Not rule-based; use custom metrics, or pipeline expectations for business rules |
| Lineage and root cause | Yes, root cause from Unity Catalog lineage |
| Alerting | Alerts on unhealthy tables and on profile metrics |
| Profiling limits | Snapshot profiles up to 4 TB; time series and inference metrics over the last 30 days by default |
| Deployment | Built into Databricks on AWS, Azure and Google Cloud; serverless compute |
| Licensing | Included in the Databricks platform; billed as serverless DBUs |
If your tables already live in Unity Catalog on Databricks, this is the lowest-effort way to get freshness and row-count anomaly detection across whole schemas, plus statistical profiling and drift tracking for important tables and ML models. It only covers Databricks, and its checks are mostly statistical, so business rules still belong in pipeline expectations or a testing tool.
Overview
Data quality monitoring is a feature of Databricks for tables governed by Unity Catalog. It has two parts. Anomaly detection is enabled per schema (or for up to 50 schemas at once from a catalog) and checks each table for freshness, flagging a table as stale when a commit is unusually late, and for completeness, flagging it when the row count over the last 24 hours falls below the expected range. It prioritises important tables through intelligent scanning and uses Unity Catalog lineage to point to a likely root cause, such as an upstream job failure.
Data profiling, the part formerly called Lakehouse Monitoring, is set up per table as a time series, snapshot or inference profile. It writes a profile metric table and a drift metric table as Delta tables in a schema you choose and creates a dashboard automatically; you can add custom metrics and alerts. Inference profiles track model inputs, predictions and performance over time. Admins also get a health view of the share of healthy tables.
Key features
- Schema-level anomaly detection for freshness and completeness, enabled with one setting
- Root cause hints from Unity Catalog lineage and health indicators on lineage graph nodes
- Data profiling with time series, snapshot and inference profiles
- Profile and drift metric tables stored as Delta tables, plus an auto-generated dashboard
- Custom metrics and alerts on profile results and on unhealthy tables
- Monitoring of ML model inputs, predictions and performance through inference profiles
- Serverless execution billed under its own DATA_QUALITY_MONITORING billing product
Integrations and ecosystem
Pros and cons
Pros
- Nothing to install: enable it on a schema or table in Unity Catalog
- Uses Unity Catalog lineage to suggest root causes
- Metric tables are ordinary Delta tables you can query and chart
- Covers ML model monitoring as well as tables
Cons
- Databricks only; no coverage of other warehouses or source systems
- Anomaly detection does not support views or foreign tables, and completeness ignores nulls, zeros and NaN values
- Snapshot profiles are limited to 4 TB and time series metrics to 30 days unless your account team changes it
- Usage cost scales with tables and frequency, and DBU rates are not shown as static values on the Databricks pricing page
Pricing
There is no separate licence: data quality monitoring needs a Databricks workspace with Unity Catalog, runs on serverless compute and is billed as serverless DBUs under the DATA_QUALITY_MONITORING billing product. Databricks says cost depends on the number and size of monitored tables and how often they are evaluated, and that anomaly detection uses intelligent scanning to limit cost. DBU rates vary by cloud, region and contract; see the Databricks profile for the billing model. Databricks Free Edition cannot be used commercially.
Pricing checked on the vendor's official page on 7 November 2026. Prices change; confirm before buying.
Who it is best for
Databricks teams with tables in Unity Catalog who want automatic freshness and volume checks across schemas, plus profiling and drift tracking for key tables and models.
Typical use cases
- Alerting when a production table stops being updated on schedule
- Catching a sudden drop in daily rows after an upstream job change
- Tracking feature drift and prediction quality of a model served on Databricks
- Giving data owners a dashboard of distribution and null statistics for a critical table
Who should look elsewhere
- Estates that span several warehouses: see Monte Carlo or Bigeye.
- Teams that need explicit business rules and tests in code: see dbt data tests or Great Expectations.
Verdict
For Databricks users, data quality monitoring gives broad freshness and completeness coverage and table profiling with almost no setup, and the lineage-based root cause is useful. It is limited to Unity Catalog tables and to statistical checks, so most teams will still pair it with explicit tests, and should watch the serverless DBU line on their bill.
Frequently asked questions
Is Lakehouse Monitoring the same as data quality monitoring?
Partly. Databricks now uses data quality monitoring as the overall name, with two components: anomaly detection and data profiling. The documentation says data profiling was formerly known as Lakehouse Monitoring.
What does it cost?
There is no separate licence. It runs on serverless compute and is billed as serverless DBUs under the DATA_QUALITY_MONITORING billing product, so cost depends on how many tables you monitor, their size and how often they are checked.
What do I need to use it?
A Unity Catalog-enabled workspace with serverless compute available for anomaly detection, and privileges such as MANAGE on the schema to enable it. Data profiling also needs access to Databricks SQL.
Sources
- Databricks: data quality monitoring
- Databricks: anomaly detection
- Databricks: data profiling
- Databricks pricing
Checked November 2026.
Listed in: Data Quality & Observability.
How we research reviews and what star ratings mean: our editorial method.