Skip to content

Scoring Methodology — Overview

This section documents how the WAF Assessment Tool scores your Databricks workspace across the four Well-Architected Framework pillars. Each pillar has its own scoring page describing the SQL logic, control thresholds, and remediation guidance.

How scoring works

The WAF Reload job (waf_reload.py) runs a set of SQL queries against Databricks system tables and stores the results in waf_cache — a catalog created in your workspace during installation.

For each control:

  1. A scoring query is executed against system tables (billing, compute, information_schema, access).
  2. The result is expressed as a percentage (0–100).
  3. The percentage is compared against a threshold defined per control.
  4. If the percentage meets or exceeds the threshold, the control is marked Met; otherwise Not Met.

Data freshness

Scores reflect the state of your workspace at the time the WAF Reload job last ran. Re-run the job from the Databricks App or the WAF Reload job directly to refresh.

Pillars

Pillar WAF ID Prefix Controls Cache Table
Data & AI Governance DG- 8 waf_controls_g
Cost Optimization CO- 8 waf_controls_c
Performance Efficiency PE- 7 waf_controls_p
Reliability R- 7 waf_controls_r

Score interpretation

Score vs Threshold Status
score >= threshold Met
score < threshold Not Met

Governance column naming

The waf_controls_g table uses a description column where other pillar tables use best_practice. This is because the governance controls were built from a different source schema. The scoring logic and Met/Not Met semantics are identical across all pillars.

Pillar pages