AI engineering glossary · Data engineering
What is data observability?
Data observability: Continuously monitoring the health of data and pipelines, using signals like freshness, volume, schema changes, distribution shifts and lineage, so problems are detected and traced before users notice. It resembles monitoring for software.
Data observability: an example
An alert fires because today's orders table has 70 percent fewer rows than usual, and lineage shows the source extract failed.
Related terms
More Data engineering terms
- Data pipeline
- Data profiling
- Data quality
- Data retention policy
- Data serialization
- Data skew
- Data SLA
- Data tiering
Where you learn it
Data observability is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.