SonarQube tells you about code smells. It doesn't tell you that your billing service is coupled to your auth service through a shared database table that nobody documented, and the one engineer who understands the relationship is about to leave.
That's technical debt. The kind that actually hurts.
Code quality tools measure:
These are useful. They're also superficial. A codebase can score A+ on SonarQube and still be an operational nightmare because the real debt is structural, not syntactic.
This is coupling that shouldn't exist. Services that depend on each other's internals. Shared database tables between services that were supposed to be independent. Feature flags that never got cleaned up and now control 15 different code paths.
How to measure it:
This is tribal knowledge that hasn't been captured. Code areas where only one person understands the behavior. Systems where the "documentation" is "ask Mike."
How to measure it:
This is the gap between how the code is structured and how it needs to be structured. A monolith that needs to be decomposed. A service boundary that's in the wrong place. An abstraction that's been patched so many times it's harder to understand than the code it was supposed to simplify.
How to measure it:
Instead of a SonarQube grade, imagine a dashboard that shows:
This is what Glue's team insights and code health features provide. Not syntax-level metrics — system-level intelligence about where your real risks are.
Technical debt metrics are only useful if they drive decisions:
The difference between "we have technical debt" (everyone knows this) and "we have 3 knowledge-concentrated features, 2 coupling anomalies, and 1 high-churn area with zero test coverage" (actionable intelligence) is the difference between complaining and fixing.
SonarQube measures syntax-level debt. The real debt — architectural complexity, knowledge concentration, dependency rot — is a form of the Understanding Tax that compounds over time.
For a specific type of hidden debt, read Feature Flags Are Technical Debt in Disguise.
Glue measures what SonarQube can't: feature-level coupling, knowledge concentration risk, and dependency anomalies. This is pre-code intelligence applied to technical debt — structural metrics derived from your actual codebase, not just syntax analysis.