Single point of failure. What if your machine needs reboot to apply kernel patches. What if your network connection breaks. Additionally, new crazy things are

Single point of failure. What if your machine needs reboot to apply kernel patches. What if your network connection breaks. Additionally, new crazy things are emerging like service mesh (Istio, Linkerd) which make observability and security easier. Of course, it all is getting quite complex. But it brings value in the end.