golang-error-handling

Idiomatic Go error handling uses %w for wrapping and errors.Is/As for inspection, applying a single rule to log or return, never both. This prevents duplicate logs and preserves error context across layers.

Use errors.Is to check for specific sentinel errors and errors.As to extract custom error types. In Go middleware, inspect wrapped errors to determine HTTP status codes while adhering to the log-or-return rule.

Structured logging with slog attaches contextual fields to errors at production boundaries. Using samber/oops, you log the error once with metadata at the top level, maintaining a clean log-or-return propagation flow.

Use Go panic and recover only for truly unrecoverable states, not for standard control flow. Safe recovery middleware catches panics to prevent application crashes while normal logic relies on error wrapping and returning.

Yes, errors.Join combines multiple errors into a single error chain. This Go feature is suitable for production applications where concurrent operations or validation steps generate multiple distinct failures.