With all the benefits that a proxy brings, you might be tempted to use them a lot—and a good idea that would be, too. Proxies are really great. There's still the issue of monitoring availability for hosts behind proxies. If a proxy goes down or cannot communicate with the Zabbix server, we would be missing data for all the hosts behind that proxy. If we used the nodata() trigger function to detect unavailable hosts (we could call such triggers availability triggers), that could mean thousands of hosts are declared unavailable. Not a desirable situation. There is no built-in dependency for hosts behind a proxy, but we can monitor proxy availability and set trigger dependencies for all hosts behind that proxy. But what should we set those dependencies to? Let's discuss the available ways to monitor proxy availability and their potential shortcomings.