AI Error Handling Generator

The system includes a base application error class with subclasses for each error type, middleware that catches and formats errors consistently, HTTP status code mapping, structured logging with error context and stack traces, user-friendly error messages separate from developer details, and optional retry logic for transient failures like network errors and timeouts.

A base AppError class extends the native Error with properties like statusCode, errorCode, and isOperational. Subclasses like ValidationError, AuthenticationError, and NotFoundError set appropriate defaults. This hierarchy lets you catch specific error types when needed or handle all application errors generically through the base class.

Errors are formatted into a consistent JSON structure with a status code, error code for programmatic handling, human-readable message, and optional field-level details for validation errors. In development, stack traces are included. In production, internal details are hidden and replaced with generic messages to prevent information leakage.

Operational errors are expected failures like validation errors, authentication failures, and network timeouts — the application knows how to handle them. Programmer errors are bugs like null references and type errors that indicate code defects. The system treats them differently: operational errors return appropriate responses while programmer errors trigger alerts and graceful recovery.

The Standard feature level includes a retry utility with exponential backoff for transient errors like database connection failures and external API timeouts. It retries the operation with increasing delays, adds jitter to prevent thundering herd problems, and gives up after a configurable number of attempts. Only transient errors trigger retries — permanent errors fail immediately.