error-handling

Standardize error handling by wrapping server route handlers with logAndSendError for 500-level failures and sendError for validation responses, while placing ErrorBoundary components around client UI sections to contain crashes and maintain graceful user-facing failures.

Return client-friendly Zod validation errors by applying the readable fromZodError messages format to generate structured 400 responses. Use the sendError helper to ensure validation failures are communicated clearly to the client without exposing internal implementation details.

Prevent fire-and-forget logging from breaking server request flow by enforcing non-blocking side effects. This approach ensures that activity logging and Sentry reporting execute asynchronously without interrupting the primary request processing, maintaining reliable and consistent application performance.

Yes, you can use a SelfHealingBoundary to handle transient async failures through automated retry logic with exponential backoff. This component automatically recovers failed UI sections without requiring a full page reload, ensuring resilient client-side error recovery.

Track and deduplicate financial calculation errors by implementing the FinancialCalculationError class, which uses deterministic input hashing to generate unique fingerprints. Sentry tag helpers contextualize these engine failures, preventing duplicate reports and improving diagnostic reliability.

Yes, this error handling approach supports graceful Sentry degradation. It is configured with environment-specific sampling and safe client reporting rules, ensuring that if Sentry operations fail, the primary application flow remains completely unaffected and stable.