zef/retry

Motivation

I first conceived and built this library while working on an app in which we would commonly encounter errors that were frustrating to users, yet recoverable by re-running certain bits of code.

This library can be used to help provide a more reliable user-experience when the services that an app relies upon have intermittent failures.

Usage

For now, see the tests for basic usage.

Logging and Metrics

When trying to diagnose the source of problems and failures in an app, it's important to understand which failure points exist, and which of those can be recovered through retrying. Therefore, this library provides a mechanism by which we can discover when retrying is worthwhile and when it's not.

This can be done through the implementing the following callbacks:

onSuccess
// Use this to gauge whether retrying is ever effective
onSuccessAfterRetry
onRetryAttempt
onFinalFailure

These can be implemented in a global location and will apply to all of the uses of Retry calls in your app. This keeps the callsites clean and introduces as little complexity as possible.

Ideas:

Add a short delay.

In some cases, you may want to insert a short delay before retrying. Depending on the source of the failure, a short delay may help the chances of a retry succeeding. I thought about including this functionality directly in the library, but have not found the need to yet myself. You could easily do this in your own function calls, and maybe even scale the delay time by attempt.currentAttempt to increase the delay of each subsequent retry before finally giving up.

Package Metadata

Repository: zef/retry

Default branch: main

README: README.md