The usual reason given for deflection control (at the established levels) is prevention of cracks. For one on-span specification, only the relative amount of deflection has meaning to this respect, i.e., if you are already respecting the L/something spec you are already complying with the more meaningful part of the specification.
The absolute statement of maximum deflection is a complementary one that of course if more stringent it is also preventing the formation of cracks (you enter a higher protection mode), but everyone can understand that the main intent is aesthetic in this case in the sense of that it is thought that above such visible deflection the bottom of the supporting element becomes unsightly. It may also mean an expected degree of maximum deflection from a tolerance viewpoint for windowmakers.
The question is if stated as enforceable, or you can subject the design to your own different criterium.
If enforceable, read with attention the code searching about:
1. Are the total loads dead and live to be considered? for example in spanish CTE typically one would only need to enter half of the live intervening loads (if you assume connection that would show even in a lintel item from cemented behaviour).
2. Typically in CTE when considering the integrity of the elements, for the more stringent serviceability check a 1/500 only DL+SDL+0.5LL hypothesis in which the loads to be entered are only those incurred AFTER the construction of the element is to be considered. Is there any clause in this intent in your code? CTE is seeking what will add to the initial status, the "live deflection", such way it is specified.
Note there are other different applicable deflection limits in CTE. I only bring these comments here to show how small caps may produce a difference.
