How do I use retry with HTTP requests in Go?

Wrap your HTTP call in a closure passed to retry.Do, as shown in the README example. Use a context for timeouts and configure strategies like backoff with jitter to handle transient network failures.

What's better for Go: retry or cenkalti/backoff?

Retry excels in interruptibility and context integration for distributed systems, while cenkalti/backoff focuses on simpler exponential backoff. Choose retry for advanced control with cancellation, or backoff for basic retry needs.

How to add jitter to backoff in retry?

Use the strategy.BackoffWithJitter function with a jitter generator, such as jitter.NormalDistribution, as illustrated in the code snippet. This helps prevent synchronization issues in distributed environments.

Does retry support circuit breakers out of the box?

No, it's designed for compatibility but requires manual integration with circuit breaker packages like breaker. You'll need to set up the breaker pattern separately to enhance fault tolerance.

How to cancel a retry operation immediately?

Cancel the context passed to retry.Do; the library respects context cancellation, allowing instant interruption of retries. This is key for responsive error handling in Go applications.

What backoff algorithms does retry include?

It supports algorithms like Fibonacci backoff, and you can define custom strategies. However, detailed examples are sparse, so refer to the strategy package documentation for implementation.

Open-Awesome

retry

MITGov5.0.0

An advanced interruptible Go library for performing actions repetitively until successful, with context integration and backoff strategies.

Visit Website GitHub

344 stars14 forks0 contributors

What is retry?

Retry is a Go library that provides an advanced interruptible mechanism to perform actions repetitively until they succeed. It solves the problem of unreliable network communications in distributed systems by offering configurable retry strategies with support for cancellation via Go contexts. The library helps developers handle transient failures gracefully without writing boilerplate retry logic.

Target Audience

Go developers building distributed systems, microservices, or network-dependent applications that require robust fault tolerance and retry capabilities.

Value Proposition

Developers choose Retry for its seamless integration with Go's context package, flexible strategy configuration (including backoff and jitter), and focus on interruptibility, making it more reliable and easier to control than basic retry implementations.

Overview

♻️ The most advanced interruptible mechanism to perform actions repetitively until successful.

Use Cases

Best For

Implementing retry logic in microservices that call external APIs
Handling transient network failures in distributed Go applications
Adding configurable backoff strategies with jitter to prevent thundering herds
Creating interruptible background tasks that respect context timeouts
Building CLI tools that need to retry commands until successful
Integrating retry mechanisms with circuit breaker patterns for fault tolerance

Not Ideal For

Projects requiring simple, non-interruptible retry loops without advanced backoff strategies
Applications where strict backward compatibility across library updates is non-negotiable
Teams seeking drop-in retry solutions with extensive, ready-to-use documentation and examples
Real-time systems with hard latency constraints where any retry delay is unacceptable

Pros & Cons

Pros

Interruptible Execution

Actions can be canceled mid-execution using Go's context package, preventing infinite loops and enabling graceful shutdowns, as demonstrated in the README example with context.WithTimeout.

Flexible Strategy Configuration

Supports configurable limits, backoff algorithms like Fibonacci, and jitter for distributed systems, reducing thundering herd problems through customizable retry behavior.

Native Context Integration

Built to work seamlessly with Go's context package for timeout and cancellation propagation, making it idiomatic for Go applications and enhancing control over retry operations.

Circuit Breaker Compatibility

Designed to integrate with breaker patterns for fault tolerance, as mentioned in the philosophy, though integration requires manual setup with packages like breaker.

Cons

Incomplete Documentation

The 'How to' section is marked as 'rewriting...', indicating gaps or outdated usage guides that can hinder quick adoption and implementation.

Breaking Changes

Major releases are not backward-compatible per SemVer, requiring careful version management and potential code refactoring for upgrades, as noted in the Integration section.

Complex Setup Overhead

Configuring advanced strategies with backoff and jitter involves multiple components and dependencies, adding complexity compared to simpler retry libraries.

Frequently Asked Questions

Related Projects

fzf

:cherry_blossom: A command-line fuzzy finder

A tool for exploring each layer in a docker image

Stars54,353

Forks1,991

Last commit7 months ago

Hub

A command-line tool that makes git easier to use with GitHub.

Stars22,957

Forks2,222

Last commit2 years ago

💥 A Lodash-style Go library based on Go 1.18+ Generics (map, filter, contains, find...)

Stars21,394

Forks946