Topics in this guide
Get the week's best data engineering content
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Advanced coverage

Once you’ve rolled out the monitors above, Metaplane will be watching for execution issues and quality problems coming out of your source systems. Issues may still arise, however, when the monitors discussed so far are too zoomed out to identify a problem only visible at a finer grain. The below will guide you through how to deploy monitors to catch these challenges.

Monitoring end user-facing tables

Monitoring of tables accessed directly by end users or by software connected to your warehouse differs from monitoring landing and transformation output tables. This type of data is frequently viewed by users after it has been aggregated into different dimensions: sales by region, click-through rate by marketing campaign, performance by product category, or a myriad of other groupings. A data quality issue affecting just one region, marketing campaign, or product category may not be large enough to trigger an alert when viewing all regions or campaigns or categories together. In order to catch data quality issues at this level, monitors must understand the different expectations of data within those varying groupings.

Metaplane offers Group By monitors to build predictive ranges for each dimension that your end users look at. Imagine a scenario where a data issue arises in a single marketing campaign, but the monitor tracking row count and mean customer purchase size across all marketing campaigns doesn’t change enough to trigger an alert.

Deploying row count and mean customer purchase Group By monitors on the marketing campaign dimension, however, would build an understanding of normal row count volumes according to each campaign (understanding that some will have much larger volumes than others). With the sensitivity to flag what is normal or anomalous for each campaign, the monitor will alert you to a stark drop off in a single campaign’s performance while all others behaved normally. Group By monitors can be deployed for any standard monitor, or they can be written into Custom SQL.

End user monitor coverage examples


Company leaders read the reports run by your analysts and analytics engineers, so use nullness Group By monitors to make sure that every value is populated. Get ahead of questions about unexpected changes in the reported KPIs with a mean Group By monitor to alert on any abrupt shifts in the data before it’s shown to leadership.

Data science

Data science teams care about model performance and data quality. Deploy nullness, cardinality, and column count monitors to ensure the data feeding their models is cleaned, and use Group By monitors to review model performance. 


Customer identity data is vital to marketing teams. Use nullness and uniqueness monitors to make sure contact information is populated and accurate as it enters your warehouse. For data that will be used for marketing activation, deploy Group By monitors to check for data quality issues in campaign performance, click-through rates, and A/B test results.


Track financial indicators like monthly recurring revenue with numeric monitors. Use Group By monitors to track if certain assets are generating less revenue than expected or to ensure a healthy flow of data from a variety of business units.

Ensure trust in data

Start monitoring your data in minutes.

Connect your warehouse and start generating a baseline in less than 10 minutes. Start for free, no credit-card required.