Skip to main content
Lineage derived from static code analysis is an approximation of your system’s real behavior. Rather than presenting a partial graph as if it were complete, Gable makes gaps explicit — so you always know which parts of the graph you can rely on and which parts still need attention.

Where gaps come from

When lineage is missing or incomplete, the cause is usually one of the following:
  • A service hasn’t been scanned yet, or is written in a language Gable doesn’t analyze yet. Static code analysis currently supports Java and TypeScript services.
  • A detection pattern is missing. Gable’s analysis looks for declared patterns of data entering and leaving a service. If a service uses an access pattern or abstraction the analysis hasn’t been configured to recognize, the corresponding lineage won’t be detected. See Understanding the Source of Lineage Issues.
  • A data store isn’t represented. Tables, streams, and files that aren’t created automatically need to be added as data store components — see Services & Data Components.
  • A cross-service connection hasn’t been made. In-service lineage only chains into end-to-end lineage when cross-service edges connect the components — see Dependencies (Cross-Service Edges).

Closing gaps

To improve coverage:
  1. Add missing components. Register data stores that participate in lineage but weren’t discovered automatically.
  2. Connect the graph. Create cross-service edges between components so field lineage flows end to end.
  3. Override incorrect or missing paths. When scanned lineage within a service is incomplete or wrong, suppress the incorrect path and create a path override — see Field Mappings & Overrides.
  4. Report detection gaps. If behavior didn’t change but lineage did — or lineage is missing for a pattern that should be detected — report a defect so Gable’s detection can be improved. These are typically one-time fixes tied to a specific access pattern.

Measuring coverage

Gable actively monitors for coverage regressions by comparing scanned output against a verified sample of known-correct field mappings. If you are interested in learning more about how coverage is measured for your services, reach out to Gable.