Network monitoring watches the fleet: which players are online, whether they are playing what they should be, and whether anything is degrading. The point is finding faults before a customer does.
The failure it exists to catch is silent. A screen showing a black rectangle, or last month's campaign, does not raise a ticket: staff stop noticing within a week, and the fault persists until somebody senior walks past. On a distributed network, the mean time to discovery without monitoring is measured in weeks.
The caveat is that "online" is a weak health check and most dashboards stop there. A player can be reachable, reporting healthy, and displaying nothing, because the render process hung, an asset failed to download, or the HDMI input switched. Useful monitoring reports what is actually on screen, ideally by periodic screenshot, and alerts on content age as well as connectivity.
Then decide who receives the alert and what they are expected to do. A monitoring system nobody is rostered to watch converts an unknown fault into a recorded one, which is not the same as fixing it.