Installation
Requirements, the go get line, and the two settings that have to match your Laravel app.
Requirements
| Version | |
|---|---|
| Go | 1.26 or newer |
| Redis | 6 or newer |
| Laravel | 10, 11, 12 or 13, using the redis queue connection |
Your Laravel app needs QUEUE_CONNECTION=redis. Either Redis client (phpredis or predis) is fine.
Install
go get github.com/rjp2525/laravel-queue-go
The package builds on go-redis. You create the Redis client yourself and hand it over, so connection pooling, TLS, Sentinel and Cluster are configured the way you already know.
import (
"github.com/redis/go-redis/v9"
"github.com/rjp2525/laravel-queue-go/pkg/queue"
)
rdb := redis.NewClient(&redis.Options{Addr: "127.0.0.1:6379"})
driver := queue.NewRedisDriver(rdb, queue.RedisDriverConfig{
Prefix: "laravel-database-",
RetryAfter: 90 * time.Second,
})
The two things that must match Laravel
Almost every "my worker starts but never sees a job" report comes down to one of these.
1. The Redis key prefix
Laravel prefixes every Redis key. The default comes from config/database.php and depends on APP_NAME and on which Laravel version generated that file:
| App created on | APP_NAME=Acme Shop gives |
|---|---|
| Laravel 10 or 11 | acme_shop_database_ |
| Laravel 12 or 13 | acme-shop-database- |
An app that started on Laravel 10 and was upgraded since still has the underscore form, because the config file is yours and upgrades don't rewrite it. Don't guess. Ask the app:
php artisan tinker --execute="echo config('database.redis.options.prefix');"
Or ask Redis, from Go:
queues, err := queue.DiscoverQueues(ctx, rdb)
// [{Prefix: "acme_shop_database_", Queue: "default"}, ...]
DiscoverQueues only finds queues that currently hold jobs, so dispatch one first.
If your app sets an empty prefix, pass NoPrefix: true. An empty Prefix on its own means "use the default".
2. retry_after
RetryAfter is how long a job may stay reserved before Redis hands it to another worker. It's the same setting as retry_after in config/queue.php (90 seconds by default), and it should be longer than your slowest job. If a job takes 5 minutes and RetryAfter is 90 seconds, a second worker will start the same job while the first is still on it.
Give slow jobs their own queue with a generous RetryAfter, rather than raising it for everything. A long RetryAfter also means a job whose worker crashed waits that long before it is retried.
Check the connection
A ten-line sanity check before you write any handlers:
size, err := driver.Size(ctx, "default")
if err != nil {
log.Fatal(err)
}
fmt.Println("jobs on default:", size)
Dispatch a job from Laravel, run that, and you should see 1. If you see 0, the prefix is wrong. Next: write your first handler.