@commy2 Looks perfect. How bad could performance suffer if allMapMarkers is checked each frame?
Maybe a slight delay could make sense here?

And what about automatic generated markers like from Blue Force Tracking etc...

How bad could performance suffer if allMapMarkers is checked each frame?

I don't think much. allMapMarkers is not like nearestObjects. It just reads an already existing array out of C++ land afaik.

Maybe a slight delay could make sense here?

maybe. although if you make the delay too big, you could miss some obviously. I guess 0.1 is fine?

And what about automatic generated markers like from Blue Force Tracking etc...

it would execute for every marker that appears in allMapMarkers

why not using a Statemachine?

There are no states here.

I'm not sure about the delay; all CBA event handlers are instantaneous as far as I know, and having these ones behaving differently might be confusing. Also, it might break some potential usages that require low latency.

It all comes down to the cost of:

private _newAllMapMarkers = allMapMarkers;
if !(_newAllMapMarkers isEqualTo GVAR(oldMarkers)) then {

have you measured the performance of that?

oldMarkers = allMapMarkers;

private _newAllMapMarkers = allMapMarkers; 
if !(_newAllMapMarkers isEqualTo oldMarkers) then {};

0 markers: 0.0036 ms
100 markers: 0.0148 ms

100 markers using private _newAllMapMarkers = allMapMarkers:
0.0137 ms

So allMapMarkers does get slower with many markers placed, but it isn't that bad. Not sure if that should run every frame though.

This are the theoretical frame decreases of running this every frame with 200 markers (assuming 0.015 x 2 = 0.03 ms):

60.0000 fps ->59.9999 fps
90.0000 fps -> 89.9997 fps
120.0000 fps -> 119.9995 fps

All in all completely unnoticeable.

reverted it :)