By “monitor” I presume you mean “determine if the (insert resource here) is available and ready to successfully perform workload at a precise point in time.”
There is no simple answer, other than attempting the intended workload. It either works, or it doesn’t. Cost-benefit analysis will dictate HA and DR efforts to mitigate failures.
The same question has been discussed in other contexts, including queue manager, channel, data base manager, specific application, and other resources.