BUYER GUIDE
Purview Data Quality Pricing and Microsoft Purview Data Quality Cost per Rule, Table and Month
Microsoft Purview bills data quality on two meters, and neither one is a price you can read off a plan card. Rules and scans consume Data Governance Processing Units, and every table you put under a data quality scan becomes a governed asset with its own daily charge. The Purview pricing page shows N/A for both until you pick a region, so most estimates stop at the definitions. This page reads the East US rates from the Azure Retail Prices API, applies the consumption figures Microsoft publishes in its own billing documentation, and turns them into a monthly number for 50, 100 and 250 tables, so you can compare Purview with a flat monitoring plan from 99 dollars a month before you enable pay-as-you-go.
14-day trial, no credit card, read-only connection
Alerted #data-eng 0.8s ago.
Downstream impact · consumers at risk
How much does Microsoft Purview Data Quality cost?
Microsoft Purview Data Quality costs 15 dollars per Data Governance Processing Unit on the default Basic SKU, 60 on Standard and 240 on Advanced, at East US list rates. Microsoft documents about 0.02 DGPU per rule per run for simple rules, so each rule run costs roughly 0.30 dollars on Basic. Every table in a data quality scan is also a governed asset at 0.0165 dollars a day, about 0.50 a month. Five rules on 100 tables once a day comes to roughly 4,550 dollars a month on Basic. DataObservability monitors up to 250 Snowflake, BigQuery, Databricks or Redshift tables for 299 dollars a month billed yearly, with no per-run charge.
Side by side
Purview Data Quality pricing compared
Swipe to see all columns →
| Charge | Rate or figure | When it applies | What it means for the bill |
|---|---|---|---|
| DGPU, Basic SKU | 15 USD per DGPU | Default for every data quality rule and health control | One DGPU is 60 minutes of managed compute |
| DGPU, Standard SKU | 60 USD per DGPU | When you switch to faster compute | Four times Basic for the same unit |
| DGPU, Advanced SKU | 240 USD per DGPU | Largest performance option | Sixteen times Basic for the same unit |
| Simple rule run | About 0.02 DGPU | Blank check, regex, unique check on 1 million rows | About 0.30 USD per rule run on Basic |
| Complex rule run | 0.03 to 0.04 DGPU | Multi-column duplicate check, table lookup, ID validation | About 0.45 to 0.60 USD per rule run on Basic |
| Governed asset | 0.0165 USD per asset per day | Any table linked to a data product, glossary term or data quality | About 0.50 USD per table per month |
| Data health controls | Basic SKU DGPUs | Daily by default once a governance domain exists | Runs on its own schedule until you disable it |
| Metadata scanning | No charge after pay-as-you-go consent | Registering and scanning sources into the Data Map | Classic Data Map customers still pay capacity units |
| Region | Rates vary by Azure region | Account and source must share a region | East US figures used on this page |
| DataObservability Starter | 99 USD a month, billed yearly | 50 tables, one warehouse | No per-rule or per-run charge, Slack alerts |
| DataObservability Team | 299 USD a month, billed yearly | 250 tables, PagerDuty, ML anomaly detection | End-to-end lineage, 90 day history |
| DataObservability Scale | 799 USD a month, billed yearly | 1,500 tables, column-level lineage, SSO | Multi-warehouse, audit log included |
Positioning and pricing models are summarized in good faith from each vendor's public pages, September 2026. Verify current terms with the vendor.
What you get
What Purview Data Quality costs, line by line
Two meters, and the pricing page shows neither number
Purview data governance has exactly two pay-as-you-go meters that matter for data quality. The first counts unique governed assets per day. The second counts Data Governance Processing Units consumed by data quality rules, profiling and health controls. When we opened the Purview pricing page in September 2026, both rows displayed N/A with the note that pricing is not available in the selected region, because the page waits for a region choice before rendering any figure. The Azure Retail Prices API, which is the public feed behind the Azure calculator, returns the East US list rates directly: 15 dollars per DGPU on Basic, 60 on Standard, 240 on Advanced, and 0.0165 dollars per governed asset per day. Those are the numbers every calculation on this page starts from.
The unit is compute time, so the bill follows rules and frequency
A DGPU is 60 minutes of fully managed compute, and Microsoft says consumption depends on three things: the rule type, the volume of data, and the source type. Its billing documentation publishes a table of real consumption. On Azure SQL Database with 1 million rows, a blank check, a regex check, a unique check and a three-column duplicate check each consume 0.02 DGPU per rule per run, and a table lookup consumes 0.03. A second table on the Standard option shows 0.02 DGPU up to 100 million records for simple rules, rising to 0.03 or 0.04 at a billion records or for complex rules. So row counts barely move the bill. The number of rules and how often you run them move it almost linearly.
Data quality turns every checked table into a governed asset
Microsoft charges the governed asset meter only for tables linked to a governance concept, and it names data quality as one of those concepts. The documented workflow for data quality starts by adding the table to a data product, then configuring a connection, profiling, writing rules and scanning the data product. In practice, every table you check is a governed asset from the day you add it. At 0.0165 dollars a day that is about 0.50 dollars per table per month, which is small next to the DGPU line but is a charge that never stops while the table stays in a data product.
The cost that sits outside both meters: writing rules
Purview Data Quality checks what someone configures. Rules apply per column, and you can assign up to 200 rules per data asset in a scan. Microsoft offers out-of-the-box rules for completeness, consistency, conformity, accuracy, freshness and uniqueness, plus AI-generated rule suggestions, but a data quality steward still has to review them, set thresholds and keep them current as schemas change. Across 20 critical tables that is a few days. Across 250 tables it is a standing role. Priced at a loaded hourly rate, that time often exceeds the DGPU bill, and it is the part of the comparison a monitoring tool that learns table baselines on its own removes.
How it works
From connected to caught
Count the tables that need checks, not the tables in the Data Map
Only tables in a data quality scan become governed assets and consume DGPUs. List the Snowflake, Databricks or BigQuery tables that feed finance, product and customer reporting. That count drives both a Purview bill and any monitoring plan.
Decide rules per table and runs per day
Five rules per table is a modest starting point: freshness, row count, one completeness check and two column rules. Multiply rules by tables by runs per day by 0.02 DGPU, then by 15 dollars on Basic. That is your planning figure before you consent to pay-as-you-go.
Run one scheduled scan for a week and read the meter
Microsoft publishes consumption ranges, not a quote. Pick ten representative tables, schedule their scan for seven days, and read the Enterprise Data Management meter in Azure Cost Management. Scale that figure, not the brochure, to your full table list.
Trial a flat plan on the same tables
Start a 14 day trial of DataObservability, connect read-only to the same warehouse, and let freshness, volume, schema and distribution monitors generate on every table. Compare what each approach caught and what each would cost over twelve months.
Worked monthly costs at East US list rates
Take the unit Microsoft publishes for simple rules, 0.02 DGPU per rule per run, and the Basic rate of 15 dollars. One rule run costs 0.30 dollars. Five rules on one table once a day cost 1.50 dollars a day, or 45 dollars over a 30 day month, plus about 0.50 dollars for the governed asset. On that reading, 50 tables checked daily come to about 2,275 dollars a month, 100 tables to about 4,550, and 250 tables to about 11,375. Hourly checks multiply the DGPU line by 24, so 20 tables with five rules each checked every hour come to about 21,600 dollars a month. On the Standard SKU every DGPU figure is four times higher. These numbers are an extrapolation of the consumption table and rates Microsoft publishes, not a quote. If your scans bill closer to one DGPU charge per table scan than per rule, the DGPU line falls toward a fifth of these figures, so 100 tables daily would land nearer 950 dollars a month. The honest range is wide, which is exactly why the one-week pilot in step three matters.
Why Microsoft's own example understates a real estate
The Purview pricing page gives one worked example: 100 data management rules and controls in a single day, each producing 0.02 DGPU on Basic, for 2 DGPU and 30 dollars that day. It reads as cheap, and for one day it is. But 100 rule runs a day is five rules on 20 tables checked once. A mid-market analytics estate with 150 tables that matter, five rules each, checked twice a day, runs 1,500 rule evaluations a day. At the same unit that is 30 DGPU, 450 dollars a day, and about 13,500 dollars a month on Basic before governed assets. The example is accurate. It is simply sized for a proof of concept.
Basic, Standard and Advanced are the same meter at different speeds
Microsoft describes the three SKUs as performance options for the same processing unit. Basic is the default for every data quality rule and health control, and data quality on on-premises sources always uses Basic. Standard costs four times as much per DGPU and Advanced sixteen times as much. The billing documentation publishes its detailed consumption table on the Standard option, where simple rules still take about a minute and 0.02 DGPU up to 100 million records. Unless your scans are missing a deadline, Basic is the setting to budget on, and switching a scan to Standard to make it finish faster quadruples its cost.
Data health controls keep running after the pilot ends
Health management controls, self-serve analytics and reports inside Purview run on the Basic SKU. Microsoft notes that the control job is scheduled daily when first provisioned and runs as long as at least one business domain exists. Consumption depends on data volume, score calculation and report generation. The job keeps billing on its schedule until you deactivate individual controls from the schedule page, so a trial that created a governance domain can leave a small daily DGPU charge behind. Put a calendar reminder on the end of any pilot to review it.
Which warehouses Purview Data Quality actually reaches
Microsoft lists data profiling and data quality scans for Azure Data Lake Storage Gen2, Azure Databricks Unity Catalog, Azure Synapse, Azure SQL Database and Managed Instance, Google BigQuery, Snowflake and Fabric. Amazon S3 is supported only through a Fabric shortcut. Amazon Redshift does not appear on the supported sources list at all. Virtual network support is not available for BigQuery. The Purview account and the data source must sit in the same Azure region, and scans authenticate only with a managed identity. A Redshift estate, or a BigQuery project in a region Purview does not serve, cannot use Purview Data Quality regardless of price.
Limits that shape a large rollout
Microsoft documents a ceiling of 200 data quality rules per data asset per scan and concurrency limits of 10 manual scans, 25 scheduled scans, 10 profiling jobs and 10 rule suggestion jobs at a time. A 250 table estate scheduled at the same hour will queue. Data quality services run on Apache Spark 3.5 with Delta Lake 3.2.1, and profiling summaries are stored in a Microsoft managed storage account in the same region as the source. Alerts go to an email alias or distribution group when a quality threshold is missed. None of this is a flaw, but each one is a design constraint a buyer should know before committing a governance program to it.
When Purview Data Quality is the right answer, stated plainly
If your organization already runs Purview for governance, your critical tables sit in Azure Databricks, Fabric or Azure SQL, and your checks are precise business rules a steward can state, Purview Data Quality is a strong fit. Scores roll up from rule to asset to data product to governance domain, which is what a chief data officer wants to report. It lives inside your Azure tenant, your Entra ID roles and your Azure bill. For a small set of tables with a few rules each it will usually cost less per month than any paid monitoring plan, including ours, and when those tables already sit in data products for governance reasons, the governed asset charge is one you pay anyway.
Where a flat monitoring plan comes out ahead
The balance flips as tables, rules and frequency grow, and it flips hardest on coverage. Purview Data Quality watches the columns someone wrote rules for, on the tables someone placed in a data product. DataObservability connects read-only to Snowflake, BigQuery, Databricks or Redshift, learns a baseline for freshness, volume, schema and distribution on every table from its own history, and alerts in Slack on every plan and PagerDuty from Team up. Starter is 99 dollars a month billed yearly for 50 tables, Team is 299 for 250 tables with end-to-end lineage and ML anomaly detection, and Scale is 799 for 1,500 tables with column-level lineage, SSO and an audit log. The price does not change when you check more often. Many teams run both: Purview for the governed business rules a steward owns, and a monitor for everything that breaks between them.
Questions buyers ask
Purview Data Quality pricing FAQ
Is Microsoft Purview Data Quality free?
No. Metadata scanning into the Data Map carries no charge once you consent to pay-as-you-go, but data quality rules, profiling and health controls consume DGPUs at 15 dollars each on Basic in East US, and every table in a data quality scan bills as a governed asset at 0.0165 dollars a day.
What is a DGPU in Microsoft Purview?
A Data Governance Processing Unit is 60 minutes of fully managed compute used by data quality and data health management. It comes in Basic, Standard and Advanced performance options at 15, 60 and 240 dollars per DGPU at East US list rates. Basic is the default for every rule.
How much does a Purview data quality rule cost per run?
Microsoft documents about 0.02 DGPU per rule per run for simple rules on 1 million rows, and 0.03 to 0.04 for complex rules or very large tables. On Basic that is roughly 0.30 to 0.60 dollars per rule run. Standard quadruples it.
How much does Purview Data Quality cost for 100 tables?
With five simple rules per table checked once a day on Basic, about 4,500 dollars a month in DGPUs plus about 50 dollars for 100 governed assets. If your scans bill per table rather than per rule, the DGPU line falls to roughly 900 dollars. Run a one-week pilot to find out which applies.
Does Purview Data Quality support Snowflake and Databricks?
Yes. Microsoft lists profiling and data quality scans for Snowflake, Azure Databricks Unity Catalog and Google BigQuery, with virtual network support for Snowflake and Databricks but not BigQuery. The Purview account and the source must be in the same Azure region.
Does Purview Data Quality work with Amazon Redshift?
Redshift is not on the list of data quality sources Microsoft publishes. Amazon S3 is supported only through a Fabric shortcut. Teams with Redshift tables need a different tool for data quality monitoring on those tables.
Why does the Purview pricing page show N/A?
The data governance rows render N/A until a region is selected, with a note that pricing is not available in the selected region. The Azure Retail Prices API and the Azure pricing calculator return the list rates for each region, including 15, 60 and 240 dollars per DGPU in East US.
Is Purview Data Quality cheaper than a data observability tool?
For a few tables, yes. Five rules on one table checked daily cost about 45 dollars a month on Basic, so the DGPU bill passes the 99 dollar Starter plan at the third table if charges run per rule, or near the tenth if they run per table scan, before counting steward time.
More of the platform
Catch broken data before your stakeholders do
Connect your warehouse and get purview data quality pricing live from one read-only connection. Transparent pricing, no credit card.