Data Quality, Announcements

Just Launched: AI Anomaly Detection For SQL Server

SQL server AI anomaly detection

Michael Segner

Michael writes about data engineering, data quality, and data teams.


Mei Tao

Product at Monte Carlo

Unlike traditional data quality solutions, Monte Carlo was originally designed to reduce data downtime across modern data platforms such as Snowflake, Databricks, Redshift, BigQuery, Azure Synapse and more. 

As we worked with data teams, we ran into a diverse set of data platforms teams used to power their data products including:

  • Postgres
  • Teradata
  • MySQL
  • Oracle
  • SQL Server

Last year we launched custom monitors, or data tests, for these environments to help identify bad data as early in the process as possible. We also launched cross database rules to help ensure data consistency across databases.

And today, we are thrilled to announce AI anomaly detection for SQL Server! 

This is a game changer for organizations that leverage one of the most popular relational databases as either one of their key data sources or as their main analytical platform. Here’s why.

AI Anomaly Detection For The Win

Today most data teams are using manual SQL queries with predefined thresholds (data tests or rules) to detect bad data in SQL Server.

AI anomaly detection is generally more efficient to deploy and produces better coverage. This is because:

  • In most cases, AI is better at setting and updating thresholds across a dataset. Manually set thresholds are noisy and create alert fatigue.
  • Writing SQL tests at scale takes time and is tedious, even when partially automated with templates. Updating and changing these tests as pipelines change is just as inefficient.

While you should take an AI-first approach to detection, you need to incorporate both strategies, which can be done entirely within the Monte Carlo platform. 

Monte Carlo is rapidly expanding our support for the Microsoft ecosystem. This advance in SQL Server monitoring follows recent announcements on Azure Synapse and Microsoft Data Fabric.

With this extensive capability across the Microsoft ecosystem, it only made sense to expand the ways we connect to it as well. Now Monte Carlo users can connect to Azure deployments using Azure Private Link. This is a private connection that provides enhanced protection against data exfiltration. More details can be found in our documentation.

The Roadmap

Monte Carlo’s vision is to be the central data observability platform covering all the data systems that are critical to delivering reliable data within the modern enterprise. 

You can expect to see data testing support for additional databases (like MotherDuck) and AI anomaly detection for databases we already support such as Teradata and others.

Our promise: we will show you the product.