p-retry

Allows adjustable delays between retries using factors and timeouts, such as minTimeout and maxTimeout, to avoid overwhelming services as detailed in the options.

Provides callbacks like shouldRetry and shouldConsumeRetry for custom logic, enabling precise error handling and retry budget management based on context.

Supports AbortSignal for clean cancellation, useful for user-initiated stops or handling process signals like SIGINT, as shown in the signal option examples.

Automatically retries on network-related TypeErrors while aborting on others, leveraging fetch specification best practices to distinguish transient failures.