Frequency is also impacted by load: the greater the load on the generator the more torque required at its input shaft to maintain the same RPM. If the generator's input engine is already at max torque then RPM must decrease all else equal. That in turn requires that every other generator on the grid also slow down to match.
When a huge chunk of generating capacity disappears there isn't enough power feeding the remaining generator input shafts (all else equal) to maintain RPM so the grid frequency must drop. That tends to destroy customer equipment among other problems.
Generators are motors and motors are generators. If the capacity disappears too quickly the grid _drives the generator as a motor_ potentially with megawatts of capacity all trying to instantly make that 100 ton rotor change from 3600 RPM to 2800 RPM or whatever. Inertia puts its $0.02 and the net result is a disintegrating rotor slinging molten metal and chunks of itself out while the bearings turn into dust.
Protective equipment sees this happening and trips the generator offline to protect it. Usually the coordinating grid entity keeps spare capacity available at all times to respond to loss of other capacity or demand changes. This is also the point of "load shedding": if spare capacity drops below a set level loads are turned off.
If spare capacity is not maintained or transmission line choke points present problems then capacity trip outs can cause progressive collapse as each generator sees excessive load, trips, and in turn pushes excess load to the next generator. If your grid control systems are well designed they can detect this from a central location and command parts of the grid to "island" into balanced chunks of load/capacity so the entire grid does not fully collapse.
Of course when you want to reconnect the islands it takes careful shifting of frequency to get them aligned before you can do that.
If all generators collapse you end up in a black start situation that requires careful staging lest more load than you expected jumps on the grid (maybe due to control devices being unpowered or stuck somewhere), triggering a secondary collapse.
Caveat: not a grid engineer so I may have gotten some of this wrong but hopefully it helps anyone who wonders why load shedding exists or how a grid can "collapse" and what the consequences are if you don't do those things and just let it ride.