App vs. DB — Who Failed?

Learning Objectives

By the end of this lesson you will be able to:

Identify whether an error originated in the application layer or the database layer
Read error signatures that distinguish the two failure sources
Write a bug report that correctly identifies the failure layer

Why the Distinction Matters

When you find a bug, your job isn't just to say "it's broken." It's to give the developer enough information to fix it efficiently.

"The page returns a 500 error" sends the developer on a hunt. "The 500 error contains a PostgreSQL 'integer out of range' exception, indicating the overflow occurs at the database INSERT layer — not in the application code" tells the developer exactly where to look.

The fix is different depending on the layer:

Application layer: validate input before passing to DB
Database layer: the column type may need to be changed (INT → BIGINT)

Recognizing Application Layer Errors

Application errors occur when the server-side code crashes before or after the database call:

HTTP/1.1 500 Internal Server Error
{"error": "ArithmeticException: / by zero",
 "trace": "at CalculationService.divide(CalculationService.java:47)"}

Signatures:

Stack trace shows Java/Python/C# class names and method names
Error type is a language-level exception (NullPointerException, ZeroDivisionError)
The database is not mentioned at all

Recognizing Database Layer Errors

Database errors occur when the application attempts a query and the DB rejects it:

HTTP/1.1 500 Internal Server Error
{"error": "DataAccessException",
 "cause": "ERROR: integer out of range",
 "detail": "Detail: Value 3000000000 is out of range for type integer"}

{"error": "PSQLException: ERROR: division by zero"}

Signatures:

Error message uses SQL or database-specific language
Mentions specific database error codes (PostgreSQL: 22003, MySQL: 1264)
Stack trace includes JDBC, JPA, Hibernate, or similar database driver classes
Error type is PSQLException, SQLException, DataAccessException

The Error Layer Decision Tree

Error received
      ↓
Does the stack trace mention Java/Python application classes?
      Yes → Application Layer
      No
      ↓
Does the error message mention SQL, PostgreSQL, integer, column, or table?
      Yes → Database Layer
      No
      ↓
      Ambiguous — report both possibilities

Common Error Signatures Quick Reference

Error Text	Layer	Meaning
`ArithmeticException: / by zero`	Application	Division by zero in app code
`NullPointerException`	Application	Null dereference in app code
`PSQLException: integer out of range`	Database	Value too large for INT column
`PSQLException: division by zero`	Database	Division by zero in SQL query
`ERROR: numeric field overflow`	Database	Value exceeds NUMERIC column precision
`DataIntegrityViolationException`	Database (via app)	Constraint violation (FK, NOT NULL, unique)

Pro Tip: When you see DataAccessException or PersistenceException in the stack trace, that's a Java wrapper around a database error. Look at the cause or detail field in the response — the actual database error is usually there.

Key Takeaways

Application errors: Java/Python exception names in stack trace, no DB mentioned
Database errors: SQL language, database exception types (PSQLException, SQLException)
The fix differs by layer: validate before DB (app fix) vs. change column type (DB fix)
Include the specific error type in your bug report — it tells the developer the layer immediately