Skip to content

go.http_retry

Stability Medium

Detects HTTP client calls without retry logic, which fail on transient network issues.

Why It Matters

Without retries:

Transient failures become permanent — Network blips cause request failures
Poor user experience — Users see errors for recoverable issues
Reduced reliability — 99.9% uptime requires handling temporary failures
Cascading issues — One failed request may abort entire workflows

Example

// ❌ Before (no retry)
func fetchUser(id string) (*User, error) {
    resp, err := http.Get(fmt.Sprintf("/users/%s", id))
    if err != nil {
        return nil, err // Fails on first transient error
    }
    // ...
}

// ✅ After (with retry)
import "github.com/hashicorp/go-retryablehttp"

var client = retryablehttp.NewClient()

func fetchUser(id string) (*User, error) {
    resp, err := client.Get(fmt.Sprintf("/users/%s", id))
    if err != nil {
        return nil, err
    }
    // ...
}

What Unfault Detects

HTTP GET/POST without retry wrapper
Missing retry on 5xx responses
Missing retry on connection errors
Direct http.Client usage without retry policy

Auto-Fix

Unfault generates patches using go-retryablehttp:

import "github.com/hashicorp/go-retryablehttp"

client := retryablehttp.NewClient()
client.RetryMax = 3
client.RetryWaitMin = 1 * time.Second
client.RetryWaitMax = 30 * time.Second