Feature request: Octane stagger the start of workers on initial server start to reduce system resource usage

[sts-ryan-holton]

Just a feature suggestion here for Laravel Octane!

Recently in a Laravel project which uses Octane with Swoole, we run 500 Octane workers via php artisan octane:start --workers=500.

On initial boot of this command, both memory usage and CPU usage peak whilst the workers boot up the application into memory and warm up everything, this means that for about 5 minutes there's system instability and performance degradation. After this, when workers are gracefully restarted or ready to handle incoming requests everything is fine.

What can we do to reduce the initial resource impact when booting all workers?

I was thinking much like Horizon where there's a "fast termination" that there should be some inverse of this for Octane? Almost like to stagger the start?