error-handling-patterns

Error handling standardization reduces failures in multi-language stacks by using language-specific patterns: Python's custom exception hierarchies and context managers, TypeScript's Result types and custom errors. Define consistent error propagation, then apply retry logic and circuit breakers for transient failures and external API calls to maintain reliability.

Exceptions interrupt control flow and suit Python and JavaScript for synchronous errors; Result types (TypeScript, Rust) make errors explicit values, enabling composition and forcing callers to handle failures. Choose exceptions for expected failures in Python, Result types for TypeScript to prevent uncaught errors in async workflows.

Retry logic with exponential backoff handles transient failures by re-attempting failed requests with increasing delays. Python supports this via custom decorators and context managers; configure retry count, initial delay, and backoff multiplier. Apply to external API calls and database operations to avoid cascading outages.

A circuit breaker prevents cascading system outages by stopping requests to failing services after a threshold, then gradually resuming traffic. Implement one for flaky databases, slow external APIs, or any dependency prone to temporary failure. Monitor state transitions and alert when circuits open.

Error aggregation collects all validation errors before returning, improving user feedback and debugging. Accumulate errors in a list during validation, then return or raise a single exception containing all failures. Apply to form validation, configuration parsing, and API request schemas.

Graceful degradation maintains partial service availability by providing fallbacks when critical features fail. Distinguish recoverable errors (retry, use cached data, reduce features) from unrecoverable ones (propagate immediately). Apply to feature flags, cached responses, and non-critical service dependencies.