Skip to content

Discuss: Component Metadata DB

Problem to solve

groupcomposition analysis needs to store a large amount of metadata about components/packages that might be used as project dependencies, to support the following features out of CI pipelines:

  • Container Scanning
  • Dependency Scanning
  • License Scanning

There's probably too much data for the primary DB. (To be further explored in #375707 (closed).)

We've explored strategies to reduce the amount of data being stored in the primary DB, but there are downsides:

Proposal

Evaluate the feasibility of a Component Metadata DB used to support product categories maintained by Composition Analysis.

Links

https://docs.gitlab.com/ee/development/database/multiple_databases.html

Edited by Fabien Catteau