Coder Social home page Coder Social logo

dotnet-retry's Introduction

DotNetRetry

Simple, lightweight retry mechanism for .NET.

Build Status NuGet contributions welcome

Table of contents

  • Description
  • Retry policies
    • Sequential
    • Exponential
  • Classes
  • Examples
  • Tests
  • License

Description

A retry library written in C#, which attempts to execute a given operation using a given retry policy. Operations can be void or can return a value. If an operation fails for any reason, a flatten AggregateException will be thrown, which contains all the exceptions occured for the method invocation.

Retry policies

Sequential

Sequential or else linear retry policy, is basically a technique which tries to execute a certain operation X amount of times with a fixed delay specified between retries. This policy will attempt to execute the given operation as many times as it is instructed to do so and will always have the same delay between these attempts. After maximum retry limit is reached, it throws a flatten AggregateException back to the caller, with all exceptions listed.

Operation supports a Forever retry mechanism, which retries indefinitely until it succeeds.

Exponential

Exponential retry policy, or exponential backoff technique, increases the wait time between retries after each failure in exponential fashion. For example, when an operation fails for the first time, the retry will be after 1 second. The second time, the retry will be after 2 seconds, after the third try, the retry will be after 4 seconds and so on and so forth. In other words, wait time increases incrementally between consecutive retry operations after each failure.

Operation supports a Forever retry mechanism, which retries indefinitely until it succeeds.

Classes

Rule

This class contains the public API for the retry attempts. It exposes a static method, Setup(Strategy strategy) that takes a Strategy enum, indicating which retry policy to use. This method returns a RuleOptions instance, which is used to setup global options for the selected retry policy. After options have been set, you can use Retry's public API to setup cancellation policies, utilize events and of course execute the operation in question.

Methods

  • RuleOptions::Setup(Strategy strategy). Public static method. Sets up the retry policy. It returns a RuleOptions class, which can be used to setup global options for specified policy.
  • void::Attempt(Action action). Executes a void synchronous Action.
  • T::Attempt<T>(Func<T> function). Executes a synchronous Func<T>.
  • Retriable::Cancel(Action<CancellationRule> cancellationRules). Specifies cancellation policies via CancellationRule instance.

Events

It also exposes events for callers to subscribe to. Events are:

  • Retriable::OnBeforeRetry(EventHandler handler). This event is dispatched before any attempt to invoke target operation.
  • Retriable::OnAfterRetry(EventHandler handler). This event is dispatched after a retry, which means after a failed but also after a successful call.
  • Retriable::OnFailure(EventHandler handler). This event is dispatched after an exception has been thrown from the target operation.

RuleOptions

Methods

  • Retriable::Config(Action<Options> options). It receives an Action<Options> object as input, to specify global options for retry policy. It returns the parent Retriable instance.

Options

A class that sets up global options for specified retry policy. Two options are included, Attempts and Time. They can be both set, or one of them.

  • int::Attempts defines the maximum number of retries. Zero indicates infinite retries.
  • TimeSpan::Time is used kind of differently in various policies. For example, Sequential policy uses this fixed time to wait after a failure before retrying again. Exponential policy uses this time as the maximum backoff time, which is used in algorithm to calculate the mean backoff time. In other words, this is the maximum wait time for each retry attempt on exponential policy.

CancellationRule

Methods

  • CancellationRule::After(TimeSpan time). This method sets up a rule to stop retrying after a specified time has passed. It returns the same CancellationRule instance, in order to chain with rest public API methods if needed.
  • ExceptionRule::OnFailure<TException>(). This method sets up a rule to stop retrying when a specific exception occurs, passing the exception as a generic parameter. It returns the same CancellationRule instance, in order to chain with rest public API methods if needed.
  • ExceptionRule::OnFailure(Type type). This method sets up a rule to stop retrying when a specific exception occurs, passing the exception as a method parameter. It returns the same CancellationRule instance, in order to chain with rest public API methods if needed.
  • IReadOnlyList<Type>::StoredExceptions returns a read-only collection with all exceptions stored as Type objects.
  • Timespan?::StoredTime is a read-only property that returns the time set by After method.

ExceptionRule

Methods

  • ExceptionRule::Or<TException>(). This is used to chain exception rules when calling OnFailure method of CancellationRule class. It adds the specified generic parameter in the exception list and cancels the retry attempts when this exception occurs.
  • ExceptionRule::Or(Type type). This is used to chain exception rules when calling OnFailure method of CancellationRule class. It adds the specified method parameter in the exception list and cancels the retry attempts when this exception occurs.
  • CancellationRule::End(). Stops the exception rules chaining and returns the parent CancellationRule instance.

Examples

Picking a policy

Use the Strategy enum

// Sets up a Sequential/Linear retry policy
var rule = Rule.SetupRules(Strategy.Sequential); 

// Sets up an expontial backoff retry policy
var rule = Rule.SetupRules(Strategy.Exponential); 

Define options

Call the Config method and provide options using the action parameter, indicating the max number of retries and/or the timing for these retries, based on the policy chosen.

// Sets up a Sequential retry policy with maximum 3 tries, having an 100 millisecond window between each retry
var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 3;
	options.Time = TimeSpan.FromMilliseconds(100);
});

// Sets up an Exponential retry policy with maximum 6 tries, having a 4 second maximum backoff time.
var rule = Rule.Setup(Strategy.Exponential).Config(options => { 
	options.Attempts = 6;
	options.Time = TimeSpan.FromSeconds(4);
});

// Sets up a Sequential retry policy with infinite tries, having an 1.5 seconds window between each retry
var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Time = TimeSpan.FromMilliseconds(1500);
});

// Sets up a Sequential retry policy with maximum 4 tries, having no window between each retry
var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 4;
});

Just a method invocation

Just retrying a method, twice, waiting for 2 seconds between retries.

var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 2;
	options.Time = TimeSpan.FromSeconds(2);
});

rule.Attempt(() => TryThisOperation());
// or just rule.Attempt(TryThisOperation);

Or passing variables

// x = 1, y = "abc"
var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 2;
	options.Time = TimeSpan.FromSeconds(2);
});

rule.Attempt(() => TryThisOperation(x, y));

Using methods that return a value

They will be treated as Func<T>

// Greet() method returns "Hello there!"
var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 2;
	options.Time = TimeSpan.FromSeconds(2);
});

var greeting = rule.Attempt(Greet);
Console.WriteLine(greeting); // Hello there!

Handling exceptions

The following will fail in all tries. If that happens, Attempt method will throw an AggregateException with all exceptions listed.

Note: Be sure to wrap your Attempt call in a try...catch block, so you can catch the flatten AggregateException.

var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 3;
	options.Time = TimeSpan.FromMilliseconds(500);
});

try 
{
    rule.Attempt(() => int.Parse("abc"));
}
catch(AggregateException ex) 
{
    // Handle all individual exceptions
}

Utilizing events

Rule class defines a public API that can utilize events. You can subscribe to certain events happening before or after a retry attempt or when a failure occurs during a retry attempt.

var rule = Rule.Setup(Strategy.Sequential).Config(options => { 
	options.Attempts = 3;
	options.Time = TimeSpan.FromMilliseconds(500);
});

rule.OnBeforeRetry((sender, args) => {
	// Do something before target method is called.
})
.OnAfterRetry((sender, args) => { 
	// Do something after target method is called.
})
.OnFailure((sender, args) => { 
	// Do something when target method has failed.
});

Cancellation rules

Rule class has available a public API which defines cancellation rules for the selected retry policy, throwing an AggregateException with all the exceptions listed up until this point. These come in three flavors:

  • Cancels the retry policy after a specific amount of time has passed.
  • Cancels the retry policy when a specific exception has occured.
  • Mix and match
// Cancels the retry policy when an ArgumentException occurs
var rule = Rule.Setup(Strategy.Sequential)
	.Config(options => { 
		options.Attempts = 3;
		options.Time = TimeSpan.FromMilliseconds(500);
	})
	.Cancel(c => c.OnFailure<ArgumentException>());

rule.Attempt(() =>
{
	// This will stop the policy from retrying this method
	// and is going to return an AggregateException including the exception below
	throw new ArgumentException("Custom Exception");
});
// Cancels the retry policy when an ArgumentException occurs with non-generic method
var rule = Rule.Setup(Strategy.Sequential)
	.Config(options => { 
		options.Attempts = 3;
		options.Time = TimeSpan.FromMilliseconds(500);
	})
	.Cancel(c => c.OnFailure(typeof(ArgumentException)));

rule.Attempt(() =>
{
	// This will stop the policy from retrying this method
	// and is going to return an AggregateException including the exception below
	throw new ArgumentException("Custom Exception");
});
// Cancels the retry policy when an ArgumentException or
// ArgumentOutOfRangeException occurs (same result with non-generic method)
var rule = Rule.Setup(Strategy.Sequential)
	.Config(options => { 
		options.Attempts = 3;
		options.Time = TimeSpan.FromMilliseconds(500);
	})
	.Cancel(c => c.OnFailure<ArgumentException>()
		      .Or<ArgumentOutOfRangeException>()
	);

rule.Attempt(() =>
{
	// This will stop the policy from retrying this method
	// and is going to return an AggregateException including the exception below
	throw new ArgumentOutOfRangeException("Custom Exception");
});
// Cancels the retry policy after 300 milliseconds passed
var rule = Rule.Setup(Strategy.Sequential)
	.Config(options => { 
		options.Attempts = 10;
		options.Time = TimeSpan.FromMilliseconds(100);
	})
	.Cancel(c => c.After(TimeSpan.FromMilliseconds(300)));

rule.Attempt(() =>
{
	// This will cause a new attempt due to failure.
	// The retry policy will cancel after 300 milliseconds have passed.
	throw new ArgumentOutOfRangeException("Custom Exception");
});
// Can mix and match cancellation rules
var rule = Rule.Setup(Strategy.Sequential)
	.Config(options => { 
		options.Attempts = 10;
		options.Time = TimeSpan.FromMilliseconds(100);
	})
	.Cancel(c => c.After(TimeSpan.FromMilliseconds(300))
	              .OnFailure<ArgumentException>());

rule.Attempt(...);

Tests

Three kinds of tests in this project

  • Unit
  • Integration
  • Acceptance

All tests are powered by the XUnit framework. Acceptance tests are using SpecFlow.

Running from command line

If you have dotnet cli installed, you can build the project from the solution folder, and then run the tests

$ > dotnet build

$ > "packages/xunit.runner.console.*/tools/xunit.console.exe" DotNetRetryTests/bin/Release/DotNetRetry.Unit.Tests.dll

$ > "packages/xunit.runner.console.*/tools/xunit.console.exe" DotNetRetry.Integration.Tests/bin/Release/DotNetRetry.Integration.Tests.dll

$ > "packages/xunit.runner.console.*/tools/xunit.console.exe" DotNetRetry.Acceptance.Tests/bin/Release/DotNetRetry.Acceptance.Tests.dll

Note: where * replace with dowloaded package

License

Library and code is free to use in commercial applications and/or libraries without restrictions.

dotnet-retry's People

Contributors

gdyrrahitis avatar

Stargazers

 avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar

dotnet-retry's Issues

Add cancellation policies

Should implement cancellation of a retry policy.
This should happen when:

  • A specific exception is met
  • Certain time has passed

Add rules to handle transient errors

Should implement rules to handle transient errors.

This should be handled by strategies. Strategies should be able to combine.
Should have strategy for web transient errors, DbProvider transient errors, File IO transient errors, etc.
Should have the ability to override/provide custom strategies.

Add event hooks

Event hooks to add

  • Hook before a retry attempt
  • Hook after a retry attempt
  • Hook on failure of target code

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.