Ploi

On Ploi, the autoscaler runs as a Daemon. The setup mirrors Forge — Ploi uses Supervisor under the hood.

1. Add the Daemon

In your Ploi site: Daemons → Add Daemon

Click Create.

If Ploi exposes advanced Supervisor options in your panel version, also set:

• stopwaitsecs = 60
• stopsignal = TERM

2. Remove competing Queue Workers

Ploi's Queue Workers panel configures separate queue:work daemons. Delete any worker whose queue the autoscaler manages, or declare those queues in queue-autoscale.excluded so the autoscaler ignores them.

See Queue Topology → Excluded Queues.

3. Deploys

Add this near the end of your deploy script, after the new release is active and migrations/config-cache steps are done:

php artisan queue:autoscale:restart

The command writes a cache signal. The running autoscale manager sees it on the next evaluation tick, gracefully stops spawned workers, and exits. Ploi's Daemon supervisor then relaunches php artisan queue:autoscale from the current release.

For manual operations, you can still use Ploi's Daemons → Restart button or restart the daemon directly with Supervisor. Find the daemon ID in the Ploi UI or via supervisorctl status.

4. Verify

In Ploi: Daemons → View Logs. First line should be something like:

[2026-04-17 12:34:56] local.INFO: Autoscale manager started

If you see a stack trace, 9 times out of 10 it's one of: missing Redis connection, missing laravel-queue-metrics publish, or a config cache from before the autoscaler was installed (php artisan config:clear).

Ploi-specific notes

• Server Cron. If you use Ploi's scheduled tasks, none of them are needed for the autoscaler — the manager is event-loop driven, not cron-driven.
• Multiple sites on one server. Each site gets its own Daemon. Don't share one autoscaler across sites — the config is per-app.
• Ploi's built-in uptime monitor will alert you if the Daemon goes down unexpectedly. That's a useful backstop in addition to in-app SLA alerts.