Alert via Laravel Log

The autoscaler already logs its own activity. This recipe adds structured, filterable alert lines to a dedicated channel so your existing log infrastructure (Papertrail, Loggly, CloudWatch, whatever) can trigger on them.

1. Add a dedicated log channel

config/logging.php:

'channels' => [
    // ... your existing channels

    'queue-alerts' => [
        'driver' => 'daily',
        'path' => storage_path('logs/queue-alerts.log'),
        'level' => 'warning',
        'days' => 14,
    ],
],

Use whatever driver fits your stack — single, syslog, papertrail, stderr for containers, etc.

2. Register the listener

app/Providers/EventServiceProvider.php:

use Cbox\LaravelQueueAutoscale\Events\SlaBreached;
use Cbox\LaravelQueueAutoscale\Events\SlaRecovered;
use Cbox\LaravelQueueAutoscale\Events\WorkersScaled;
use App\Listeners\LogQueueAlerts;

protected $listen = [
    SlaBreached::class => [LogQueueAlerts::class.'@onBreach'],
    SlaRecovered::class => [LogQueueAlerts::class.'@onRecovery'],
    WorkersScaled::class => [LogQueueAlerts::class.'@onScaled'],
];

3. The listener

app/Listeners/LogQueueAlerts.php:

<?php

namespace App\Listeners;

use Cbox\LaravelQueueAutoscale\Alerting\AlertRateLimiter;
use Cbox\LaravelQueueAutoscale\Events\SlaBreached;
use Cbox\LaravelQueueAutoscale\Events\SlaRecovered;
use Cbox\LaravelQueueAutoscale\Events\WorkersScaled;
use Illuminate\Support\Facades\Log;

class LogQueueAlerts
{
    public function __construct(private AlertRateLimiter $limiter) {}

    public function onBreach(SlaBreached $event): void
    {
        Log::channel('queue-alerts')->error('Queue SLA breach', [
            'connection' => $event->connection,
            'queue' => $event->queue,
            'oldest_job_age_seconds' => $event->oldestJobAge,
            'sla_target_seconds' => $event->slaTarget,
            'pending_jobs' => $event->pending,
            'active_workers' => $event->activeWorkers,
        ]);
    }

    public function onRecovery(SlaRecovered $event): void
    {
        Log::channel('queue-alerts')->info('Queue SLA recovered', [
            'connection' => $event->connection,
            'queue' => $event->queue,
            'current_job_age_seconds' => $event->currentJobAge,
            'sla_target_seconds' => $event->slaTarget,
        ]);
    }

    public function onScaled(WorkersScaled $event): void
    {
        // WorkersScaled fires per action — rate-limit so flapping doesn't spam the log.
        $key = "scaled:{$event->connection}:{$event->queue}";

        if (! $this->limiter->allow($key)) {
            return;
        }

        Log::channel('queue-alerts')->notice('Workers scaled', [
            'connection' => $event->connection,
            'queue' => $event->queue,
            'from' => $event->from,
            'to' => $event->to,
            'action' => $event->action,
            'reason' => $event->reason,
        ]);
    }
}

Done

Tail the log to verify:

tail -f storage/logs/queue-alerts-*.log

Point your log aggregator's alert rules at the Queue SLA breach message or the queue-alerts channel directly.

Tuning

• SlaBreached and SlaRecovered fire once per state transition, so they don't need rate-limiting. Only WorkersScaled, ScalingDecisionMade, and SlaBreachPredicted can be chatty.
• Change the cooldown in config/queue-autoscale.php → alerting.cooldown_seconds or via QUEUE_AUTOSCALE_ALERT_COOLDOWN.
• If your aggregator needs JSON, swap the daily driver for one that wraps logs in JSON formatters (e.g. papertrail or a custom Monolog processor).