Glossary
Schema Drift
Schema drift is when the structure or meaning of data changes without the downstream systems that depend on it being updated. The pipeline keeps running. The dashboards stay green. The data is wrong.
What is schema drift?
Schema drift is when the structure, format, or meaning of data in a source system changes over time without those changes being communicated to or handled by the downstream systems that consume the data. The integrations keep running and report success, but the data flowing through is now incomplete or incorrect.
What are the types of schema drift?
There are three: structural drift (a field is added, removed, renamed, or retyped), semantic drift (a field's meaning changes while its name and type stay the same), and temporal drift (the timing or ordering of data changes, such as a timezone or batch-frequency shift).
Why don't monitoring tools catch schema drift?
Most monitoring tools measure infrastructure signals — uptime, error rates, job success — not data correctness. A pipeline that drops or misreads records due to drift still reports a successful execution, so structural and semantic drift pass through undetected.
How do you detect schema drift?
Effective detection works at the semantic layer: continuously comparing schemas across integration points, baselining the expected values of each field, and recognizing when a field's meaning has shifted even if its structure has not. Techniques like vector-embedding comparison and statistical distribution monitoring make this tractable.
For a deeper treatment, read what is schema drift (and why your monitoring tools miss it) or see how mmune detects it.