Go Concurrency & Worker Pool Patterns

1. Structured Concurrency with golang.org/x/sync/errgroup

Always bound goroutine lifecycles and propagate the first encountered error:

package main

import (
    "context"
    "fmt"
    "golang.org/x/sync/errgroup"
    "time"
)

func FetchAllUserData(ctx context.Context, userIDs []string) error {
    // Cancel sibling operations immediately if any one fails
    g, ctx := errgroup.WithContext(ctx)
    g.SetLimit(10) // Bounded concurrency limit

    for _, id := range userIDs {
        userID := id // Pin loop variable (prior to Go 1.22)
        g.Go(func() error {
            return fetchProfile(ctx, userID)
        })
    }

    return g.Wait()
}

func fetchProfile(ctx context.Context, id string) error {
    select {
    case <-ctx.Done():
        return ctx.Err()
    case <-time.After(100 * time.Millisecond):
        fmt.Printf("Fetched user %s\n", id)
        return nil
    }
}

2. Goroutine Leak Prevention

• Never launch a goroutine without knowing how and when it terminates.
• Always check ctx.Done() when reading/writing unbuffered channels or waiting on IO.
• Always run the race detector in tests: go test -race ./....