[Feature Request] Improve scoring transparency and resource-level visibility for PE-02-02 and PE-02-04

[albertwangnz]

Description

While reviewing the results produced by Databricks-WAF-Light-Tooling, I noticed that some performance-related PE-02 recommendations provide only high-level conclusions, without enough transparency into how scores are calculated or which resources are being evaluated.

This limits the ability to understand the results and take targeted remediation actions.

Observed behavior

For the following recommendations:

• PE-02-02 – Design workloads for performance – Use parallel computation where it is beneficial
• PE-02-04 – Design workloads for performance – Prefer larger clusters

The tool currently provides:

• Only high-level guidance
• No explanation of the score calculation logic
• No visibility into which Databricks resources (notebooks, jobs, SQL queries, pipelines, etc.) are evaluated
• No list of resources considered “Not Met”, even when scores are low or partial

As a result, it is unclear which workloads are driving the score and what concrete changes would improve compliance.

Expected behavior / Enhancement request

Enhance these recommendations by providing:

• A clear explanation of the scoring methodology (rules, thresholds, heuristics, or signals used)
• Information about which resource types are analyzed
• A list of specific resources that are evaluated as “Not Met”, where applicable
• More actionable context to help users prioritize performance optimization work

Impact

Improved transparency would:

• Increase confidence in the reported WAF scores
• Make the assessment results more actionable
• Help engineering teams prioritize and track performance improvements aligned with WAF guidance

[albertwangnz]

From: Abhishek Pratap Singh

For PE-02-02 and PE-02-04, you're right that the current guidance is high-level. We're working on adding more transparency into the scoring methodology — in the meantime, we've added a full Scoring Methodology section to the documentation that explains the exact SQL and thresholds used for every control.