Networks don't usually crash unless a switch, router, or proxy server goes down. If a switch fails, then all the systems connected to it will not be accessible and they will not be able to access the rest of the network or the Internet. You can install more than one network interface on your servers and have each connected to a different switch to eliminate this point of failure. If a router fails, that is a bigger problem. If you have set up multiple network interfaces (NICs) on your servers, then each can be routed separately, resulting in redundant routing services, solving this problem also. Proxy servers likewise can have backup systems to take over if they fail, and your NIC's can be directed to different proxys if needed.
In any case, it is better to be proactive and monitor your network services to be sure they are healthy, and if they start to degrade in performance, schedule a replacement before they fail. This is not a simple issue. It was what I was doing at Nokia before the Microsoft take-over, developing tools to perform predictive analytics to determine when systems and services were starting to fail, but before they would fail.