Python Errors and Reliability

Overview

Design errors so callers can act and operators can diagnose quickly. Treat these recommendations as preferred defaults — when a default conflicts with project constraints, suggest a better-fit alternative and call out tradeoffs and compensating controls.

When to Use

• Implementing or reviewing timeout, deadline, or retry logic.
• Translating exceptions across layer boundaries (e.g., infra → domain).
• Classifying failures as retryable vs. permanent.
• Adding idempotency guarantees to retried writes.
• Diagnosing swallowed exceptions or silent failures in existing code.

When NOT to use:

• Pure data-transformation code with no I/O or failure modes.
• Simple validation that raises immediately with no translation needed.