m13-domain-error

Design error handling by categorizing who handles each error and how they recover. Layer 2: Design Choices Who needs to handle this error, and how should they recover? Before designing error types: Before designing error types: Who sees this error? Can we recover? What context is needed? To domain constraints (Layer 3): To implementation (Layer 1): - Distinguish between user-facing errors (actionable messages), internal errors (debug details), system errors (monitoring), and transient vs. permanent failures to determine recovery strategy - Use typed error enums with thiserror and implement is_retryable() checks to enable appropriate handling patterns - Apply recovery strategies: retry with exponential backoff for transient failures, fallback values for degraded modes, circuit breakers for cascading failures, and timeouts for slow operations - Include error context via .context() for debugging and structured logging with request IDs; avoid exposing internal details to end users - Is this user-facing or internal? - Is recovery possible? - What context is needed for debugging?