litestar-exception-handling（litestar 异常处理）

Exception Handling

Execution Workflow

1. Separate startup and configuration failures from request-time failures.
2. Decide whether the failure should be represented by raising HTTPException or by mapping another exception type through an exception handler.
3. Choose the narrowest layer that should own the handler: app, router, controller, or route handler.
4. Keep the outbound error contract stable: status code, payload shape, media type, and safe detail level.
5. Register app-level handlers for 404 Not Found and 405 Method Not Allowed when those responses must be customized.
6. Verify validation and server-failure paths separately so sensitive details do not leak unintentionally.

Core Rules

• Distinguish configuration/startup exceptions from request-handling exceptions.
• Use HTTPException or its subclasses when the handler should raise an HTTP-aware error directly.
• Use exception handlers to translate domain or library exceptions into transport-safe responses.
• Keep error payload schemas stable across handlers and layers.
• Keep documented error behavior aligned with litestar-openapi route metadata and security docs.
• Redact or replace validation and internal error details when exposing raw messages would leak implementation details.
• Prefer specific exception mappings over broad catch-alls that hide root causes.
• Keep handler registration close to the layer that owns the behavior, except for 404 and 405, which must be handled at app scope.

Decision Guide

• Raise HTTPException subclasses directly for clear HTTP semantics such as 401, 403, 404, 429, or 503.
• Map domain exceptions with exception_handlers when business logic should stay transport-agnostic.
• Map by exception class when the exception type is the stable contract.
• Map by status code when multiple exceptions should share one response format for the same HTTP status.
• Use app-level defaults for generic 500 handling and lower-level overrides only where behavior genuinely differs.

Reference Files

Read only the sections you need:

• For exception taxonomy, built-in subclasses, HTTPException constructor behavior, validation warnings, headers, extra, and websocket-related exceptions, read references/built-in-exceptions.md.
• For handler registration patterns, plain-text vs JSON responses, status-code mappings, layered overrides, and the app-only 404/405 rule, read references/handler-patterns.md.

Recommended Defaults

• Standardize one JSON error envelope unless a route intentionally serves another media type.
• Treat ValidationException.extra as potentially user-visible diagnostic data.
• Keep 500 responses generic for clients and detailed in logs.
• Prefer class-based mappings for business exceptions and app-level status-code mappings for generic fallback behavior.
• Keep middleware and dependency failures covered by the same top-level error strategy as route handlers.

Anti-Patterns

• Returning raw exception text for internal server errors in production APIs.
• Exposing validation internals or stack-trace-like details to clients without intent.
• Registering a catch-all handler that collapses every failure into the same 400 response.
• Splitting identical error envelopes across many handlers without a clear reason.
• Trying to customize 404 or 405 below the app layer.
• Encoding authorization policy inside generic exception handlers instead of security layers.

Validation Checklist

• Confirm expected exception classes map to the intended status codes.
• Confirm app startup and configuration errors are not treated as request-time API responses.
• Confirm 404 and 405 customization is registered on the Litestar app instance.
• Confirm payload shape is consistent across handlers, routes, and layers.
• Confirm documented error responses and raised exceptions match the real contract.
• Confirm validation failures do not leak more detail than intended.
• Confirm unhandled exceptions still surface as 500 responses and remain observable in logs.
• Confirm route-specific overrides do not accidentally shadow broader app policy.
• Confirm middleware and dependency exceptions follow the same contract as handler-raised exceptions.

Cross-Skill Handoffs

• Use litestar-authentication or litestar-security for auth-specific 401 and 403 strategy.
• Use litestar-responses when the main task is response formatting rather than exception mapping.
• Use litestar-testing for exhaustive failure-path assertions and override coverage.
• Use litestar-openapi when documented error responses or security-related failure docs need to match runtime behavior.
• Use litestar-logging when exception observability, redaction, or structured logging is part of the task.

Litestar References

• https://docs.litestar.dev/latest/usage/exceptions.html
• https://docs.litestar.dev/latest/reference/exceptions.html
• https://docs.litestar.dev/latest/usage/middleware/using-middleware.html