What problem does it solve?
It prevents silent failures and confusing bugs by enforcing consistent error modeling, classification, wrapping, retry logic, timeouts, and boundary logging.
Core Features & Use Cases
- Error contracts that callers can reason about: use value-based errors (e.g., Go error, Rust Result, TypeScript typed errors) and reserve exceptions/panics for programmer errors.
- Correct handling strategies via classification: distinguish transient, permanent, and programmer errors to decide whether to retry, fail fast, or crash.
- Context-preserving error chains: wrap and propagate original causes so downstream diagnostics remain accurate.
- Retry/backoff with safety rails: apply exponential backoff with jitter, enforce retry budgets, and ensure idempotency before retrying.
- Deadline-aware reliability: propagate timeouts/deadlines correctly and return distinct timeout/deadline exceeded failures.
- Boundary logging discipline: log once at the layer that handles the error, and never log from libraries.
- Resilience patterns where needed: use circuit breakers and bulkheads to prevent cascading failures instead of over-retrying.
Quick Start
Ask an AI to review your function or API design for error classification, context-wrapping, retry/idempotency safety, timeout/deadline propagation, and log-once boundary placement for your target language.
