attemptify

TypeScript retry library with no dependencies.

Objective

The aim of this module is to simplify the setup and execution of retries. When setting up a retry, there are a number of considerations, such as how long to wait, how many attempts to make, and whether to fail. This module provides a unified procedure for setting up and executing these complicated retries.

It is particularly influenced by spring-retry and resilience4j, and can incorporate retry settings in a robust and concise manner by combining objects that represent settings.

Usage

Getting started

Try to start from SimpleRetryPolicy to call ofDefaults and boot Attempt with that policy.

const simpleRetry = SimpleRetryPolicy.ofDefaults();

new Attempt(simpleRetry).execute(() => {
  return new Application().run();
});

This retry policy has settings that wait for 1 seconds in interval and max attempt count is 5 times. Attempt can execute synchronous or asynchronous, and when asynchronously want to attempt, it is possible to call executeAsync like below:

new Attempt(simpleRetry).executeAsync(() => (new Application().run()));

Attempt enable log for debug, using enableDebugLogging:

new Attempt(simpleRetry)
    .enableDebugLogging()
    .executeAsync(() => (new Application().run()));

About retry policies

This module provides some retry policies - the policy is the strategy for retry.

• SimpleRetryPolicy ... simply retrying tomax attempts
• ExponentialBackOffRetryPolicy ... wait interval increases exponentialy
• InfiniteRetryPolicy ... retry infinitely until an attempt become success

Customize retry policy

This project provides 2 retry policies: constant retry and exponential backoff. User can use SimpleRetryPolicy and ExponentialBackOffRetryPolicy each.

SimpleRetryPolicy

This policy attempt to do a function with constant interval set with interval duration and max attempt counts. RetryPolicy can customize its policy setting to set some values to constructors:

const simpleRetry = new SimpleRetryPolicy(seconds(1), 3)

The 1st argument of SimpleRetryPolicy receive Duration instance, passing easily to use msecs / seconds / minutes .

Also, if we want to set a part of settings of policy, a builder is useful. Values that is not set will be defaults.

SimpleRetryPolicy.newBuilder()
      .duration(seconds(1))
      .build();

For all builder settings, see more: Class Builder

ExponentialBackOffRetryPolicy

ExponentialBackOffRetryPolicy also can customize its retry settings by its constructor or builder.

1. initial delay ... initial interval to wait a next attempt
2. max Attempts ... max retry attempt counts
3. multiplier ... an index that multiply a previous delay

And it can be built like below:

const policy = new ExponentialBackOffRetryPolicy(seconds(1), 4, multiplierOf(2));

The 3rd parameter multiplier will multiply its interval exponentialy. For example, when interval duration is 1 seconds and multiplier is 2, interval is calcurated like 1(sec)^2 .

Like SimpleRetryPolicy, ExponentialBackOffRetryPolicy also can be built by builder style.

const policy = ExponentialBackOffRetryPolicy.newBuilder()
      .maxAttempts(2)
      .build();

InfiniteRetryPolicy

This has no builder, so build it calling constructor or ofDefaults.

For all builder settings, see more: Class Builder

Useful helper type and functions

This project provides some useful helper type and functions.

Duration

Duration class represents some time. It can be constructed by constructor or factory method inside of it. This class has time value itself and duration unit. For retrying, default duration unit is milli seconds.

// Duration unit: milliseconds
Duration.of(1000);

// Duration unit: seconds
Duration.ofSeconds(1);

// Duration unit: minutes
Duration.ofMinutes(1);

User also can use some helper functions to shortcut:

msecs(1000);

seconds(1);

minutes(1);

Multiplier

Multiplier object represents an index to multiply an interval, so this force to round to integer.

User can build it by constructor or helper functions like below:

// Both are Okay.
new Multiplier(2);

multiplierOf(2);

docs

All class documents can be seen here: https://simonnozaki.github.io/attemptify/