— try to optimise your logic first (are there any dumb things you tell the game to do that are simply a waste of time? The case jayderyu made with disabling behaviours would be just that, if you game does not require that you update the positions of the ennemies that are far away, might as well "put them to sleep")
Then try to see if there is something that is draining your performances down, if easily spotted, correct it without influencing too much your game if possible, if not spotted that easily, try to experiment and see.
However micro-optimisation is not worth the effort when it comes down to fighting with the engine (if a flappy bird game doesn't run well because "not optimised enough", don't optimise it, changing engine would be a wiser idea in the long term, perhaps even mid term run, and since you have done a design work on your game probably, porting it over will hopefully not be as bad as it sounds), not saying it will not work though if you want it really.
To be honest, micro-optimisation is something the engine should already be taking care off, if you feel it is not doing it's job properly, try to see for other alternatives, C2 is a web-based runtime game engine, any other use is not warranted to be perfect.
Tl;dr try to see if the logic is flawed first, or incomplete ("we did not agreed on the fact ennemies should be doing things far off the screen or not, the game design doc we did said nothing about that nor the requirements, so what should we do?"), then profile the game and see what is bringing it to the knees, then see if you can work to make that better or if using another engine is a better choice.
Keep in mind that this advice is applicable because of the fact C2 adds a pretty visible overhead due to its browser based nature, and this overhead can be tricky as, while great games can be done on a very limited system when we know it, here the knowledge of the system is not easy to understand as the browser layer can be really nasty or even not related to the hardware.