When the check engine light in your car turns on, all you know at that point is that you have a problem. You don’t know what happened to your engine, why it happened, how bad the issue is, or what to do about it yet.
When you take it to an auto shop, the mechanic starts the diagnostic process of finding the answers to these questions and ultimately getting you back on the road.
When an Expectation fails, the real question data teams ask next is obvious: how bad is this, really? Is this a small anomaly that will resolve itself, or a serious issue that needs immediate investigation?
That’s why we’ve shipped root cause analysis improvements in GX Cloud, focused on helping you quickly understand to what extent an Expectation failed and start fixing your data pipelines.
Know where to start troubleshooting at a glance
When you see an Expectation failure in GX Cloud, you might want to know how far an observed value is beyond its preset threshold, how many rows actually failed, or how severe the failure is.
Now you can answer these questions right in GX Cloud.
Clear context where you need it
GX Cloud now allows you to specify the verbosity of an Expectation’s validation results. Result formats can be set at the suite, checkpoint, or Asset level. You control how much detail you see before jumping into root cause analysis.
Three Levels of Insight into Expectation Failures
When an Expectation fails, you can now choose from three result formats to get quick insights:
Status only
See the status and the percentage of failed rows for Expectations that have a “mostly” parameter. GX Cloud also provides investigation-ready queries without exposing data in the UI so you can start your diagnosing right away. Perfect for respecting privacy requirements.
Observed values (default)
In addition to status, return observed values for aggregate Expectations (min, max), and sample missing or unexpected values in set Expectations. This quickly allows you to understand to what degree the Expectation has failed, without sharing data with GX beyond what is needed for computing failure.
Sample unexpected rows
For deeper investigation, view a small sample of failed rows in failing Expectations to quickly see if there’s an obvious pattern in the column under test or in the surrounding columns, allowing you to identify the root cause more quickly.
Each option is designed to answer the same core question: to what extent did this Expectation fail?
GX Cloud also renders results appropriately for different Expectation types, including aggregate, row-level, custom SQL, and multi-source Expectations.
Why This Matters
Not all data quality issues are the same, nor should your response to them be.
GX Cloud’s root cause analysis improvements help complete the data quality lifecycle in practice, helping you triage issues faster, determine their severity, and address them right away.
Ready to dig deeper into data quality? Try GX Cloud for yourself.
