Skip to content

Queue Monitor

What it monitors

  • Depth – number of pending jobs in each queue
  • Failed count – rows for that queue in Laravel’s failed_jobs table
  • Oldest job age – how long the oldest waiting job has been sitting

Supported drivers

  • redis – queries queue lengths from Redis directly
  • database – queries the jobs and failed_jobs tables

Configuration

'queues' => [
'enabled' => true,
'watch' => [], // empty = auto-discover all queues
'connections' => [], // queues on other connections (v0.9.9+)
'depth_alert_threshold' => 1000,
'wait_time_alert_seconds' => 300,
],

To watch specific queues only:

'watch' => ['default', 'emails', 'notifications'],

From package v0.9.9, queues.connections reports queues on connections other than the default one, as ['connection' => ['queue', ...]]. An empty list for a connection discovers its queues. These queues are reported as connection:queue. See the configuration reference.

Alert conditions

In the package, depth_alert_threshold and wait_time_alert_seconds drive local alerts only, which are sent when crontinel:check runs. They are not sent to the hosted app; the hosted app uses the per-app settings below.

Per-app queue thresholds (SaaS)

On Crontinel SaaS, the app settings page has one depth setting, Queue depth (jobs), default 1000. It applies to every queue in that app.

The same settings page has three optional thresholds. Each is off while blank, so an app gets no new alerts until you set one.

SettingAlerts when
Oldest job age (seconds)The oldest waiting job in any queue is older than this
Failed jobs (count)A queue’s failed job count is above this
Horizon failures per minuteHorizon’s failed jobs per minute is above this

Each alert resolves when the value drops back under its threshold. Queue and Horizon numbers come from the Laravel package.