Here :
https://github.com/api-platform/api-platform/blob/main/helm/api-platform/templates/deployment.yaml#L120-L123

lifecycle:
  preStop:
    exec:
      command: ["/bin/sh", "-c", "/bin/sleep 1; kill -QUIT 1"]

The Helm chart defines a custom preStop lifecycle hook that explicitly sends a SIGQUIT signal to PID 1 (FrankenPHP / Caddy) to stop the service.

Is this still relevant today?

Kubernetes already sends a SIGTERM to the container during pod termination and handles endpoint removal and graceful shutdown. For FrankenPHP (Caddy), SIGTERM is natively supported and should be sufficient for a clean shutdown. (https://caddyserver.com/docs/command-line#signals)

Is there still a reason to force a SIGQUIT here, or is this legacy behavior from older php-fpm–based setups?