QA (public) » openQA Project (public)

Motivation¶

As a SLE maintenance test reviewer I would like an AI chat prompt in openQA to give me back a specific combination of openQA test failures based on search parameters using human language so that I don't need to access the raw SQL database and learn SQL

Acceptance criteria¶

• AC1: openQA returns a corresponding /tests/overview URL showing results based on human language requests in an AI chat prompt accessible over the openQA webUI using common search parameters like "test name", "result", "version", "arch", "machine"
• AC2: If there are no matching results the AI chat prompt returns useful pointers to documentation, examples, etc.
• AC3: No significant detrimental performance impact on the openQA webUI
• AC4: It's still possible to access openQA results without needing to use an AI chat prompt
• AC5: We have access to a usage counter of the chat prompt

Suggestions¶

• Potentially extract a spike solution first, e.g. "[spike][timeboxed:10h] AI chat prompt in openQA returning (any) test results based on request" with goals
  • G1: We know the impact on necessary changes to the database structure needed for AI chat prompt queries
  • G2: The performance impact on running a local AI component was measured